Authors: Hong kyu Lee, Ruixuan Liu, Li Xiong
Abstract: Machine unlearning is an emerging technique that removes the influence of a subset of training data (forget set) from a model without full retraining, with applications including privacy protection, content moderation, and model correction. The key challenge lies in ensuring that the model completely forgets the knowledge of the forget set without compromising its overall utility. Existing unlearning methods for large language models (LLMs) often utilize auxiliary language models, retain datasets, or even commercial AI services for effective unlearning and maintaining the model utility. However, dependence on these external resources is often impractical and could potentially introduce additional privacy risks. In this work, we propose direct token optimization (DTO), a novel self-contained unlearning approach for LLMs that directly optimizes the token level objectives and eliminates the need for external resources. Given a sequence to unlearn, we identify two categories of tokens: target tokens, which capture critical knowledge for unlearning, and the remaining non-target tokens, which are crucial for maintaining the model utility. The former are used to optimize the unlearning objective, while the latter serve to preserve the model's performance. The experimental results show that the proposed DTO achieves up to 16.8$\times$ improvement in forget quality on several benchmark datasets than the latest baselines while maintaining a comparable level of model utility.
Authors: Kimihiro Hasegawa, Wiradee Imrattanatrai, Masaki Asada, Ken Fukuda, Teruko Mitamura
Abstract: Procedural activity assistants potentially support humans in a variety of settings, from our daily lives, e.g., cooking or assembling flat-pack furniture, to professional situations, e.g., manufacturing or biological experiments. Despite its potential use cases, the system development tailored for such an assistant is still underexplored. In this paper, we propose a novel framework, called TAMA, a Tool-Augmented Multimodal Agent, for procedural activity understanding. TAMA enables interleaved multimodal reasoning by making use of multimedia-returning tools in a training-free setting. Our experimental result on the multimodal procedural QA dataset, ProMQA-Assembly, shows that our approach can improve the performance of vision-language models, especially GPT-5 and MiMo-VL. Furthermore, our ablation studies provide empirical support for the effectiveness of two features that characterize our framework, multimedia-returning tools and agentic flexible tool selection. We believe our proposed framework and experimental results facilitate the thinking with images paradigm for video and multimodal tasks, let alone the development of procedural activity assistants.
Authors: Amirhossein Abaskohi, Tianyi Chen, Miguel Mu\~noz-M\'armol, Curtis Fox, Amrutha Varshini Ramesh, \'Etienne Marcotte, Xing Han L\`u, Nicolas Chapados, Spandana Gella, Christopher Pal, Alexandre Drouin, Issam H. Laradji
Abstract: We introduce DRBench, a benchmark for evaluating AI agents on complex, open-ended deep research tasks in enterprise settings. Unlike prior benchmarks that focus on simple questions or web-only queries, DRBench evaluates agents on multi-step queries (for example, ``What changes should we make to our product roadmap to ensure compliance with this standard?") that require identifying supporting facts from both the public web and private company knowledge base. Each task is grounded in realistic user personas and enterprise context, spanning a heterogeneous search space that includes productivity software, cloud file systems, emails, chat conversations, and the open web. Tasks are generated through a carefully designed synthesis pipeline with human-in-the-loop verification, and agents are evaluated on their ability to recall relevant insights, maintain factual accuracy, and produce coherent, well-structured reports. We release 15 deep research tasks across 10 domains, such as Sales, Cybersecurity, and Compliance. We demonstrate the effectiveness of DRBench by evaluating diverse DR agents across open- and closed-source models (such as GPT, Llama, and Qwen) and DR strategies, highlighting their strengths, weaknesses, and the critical path for advancing enterprise deep research. Code is available at https://github.com/ServiceNow/drbench.
Authors: Rik Koncel-Kedziorski, Brihi Joshi, Tim Paek
Abstract: As the adoption of language models advances, so does the need to better represent individual users to the model. Are there aspects of an individual's belief system that a language model can utilize for improved alignment? Following prior research, we investigate this question in the domain of opinion prediction by developing PrimeX, a dataset of public opinion survey data from 858 US residents with two additional sources of belief information: written explanations from the respondents for why they hold specific opinions, and the Primal World Belief survey for assessing respondent worldview. We provide an extensive initial analysis of our data and show the value of belief explanations and worldview for personalizing language models. Our results demonstrate how the additional belief information in PrimeX can benefit both the NLP and psychological research communities, opening up avenues for further study.
Authors: Shuyue Stella Li, Avinandan Bose, Faeze Brahman, Simon Shaolei Du, Pang Wei Koh, Maryam Fazel, Yulia Tsvetkov
Abstract: Current large language model (LLM) development treats task-solving and preference alignment as separate challenges, optimizing first for objective correctness, then for alignment to aggregated human preferences. This paradigm fails in human-facing applications where solving a problem correctly is insufficient if the response mismatches the user's needs. This challenge intensifies in just-in-time scenarios where no prior user interaction history exists due to cold-start conditions or privacy constraints. LLMs need to identify what they don't know about user preferences, strategically elicit preference values through questioning, then adapt their reasoning processes and responses accordingly -- a complicated chain of cognitive processes which we term personalized reasoning. We introduce PREFDISCO, an evaluation methodology that transforms static benchmarks into interactive personalization tasks using psychologically-grounded personas with sparse preferences. Our framework creates scenarios where identical questions require different reasoning chains depending on user context, as optimal explanation approaches vary by individual expertise and preferences while maintaining factual accuracy. Evaluation of 21 frontier models across 10 tasks reveals 29.0% of naive personalization attempts produce worse preference alignment than generic responses, yet generic responses also fail to serve individual user needs effectively. These findings suggest personalized reasoning requires dedicated development rather than emerging naturally. PREFDISCO establishes personalized reasoning as a measurable research frontier and reveals fundamental limitations in current LLMs' interactive capabilities, providing a foundation for developing systems that can adapt to individual users in education, healthcare, and technical domains where personalization is critical.
Authors: Xin Xu, Xunzhi He, Churan Zhi, Ruizhe Chen, Julian McAuley, Zexue He
Abstract: Existing studies on bias mitigation methods for large language models (LLMs) use diverse baselines and metrics to evaluate debiasing performance, leading to inconsistent comparisons among them. Moreover, their evaluations are mostly based on the comparison between LLMs' probabilities of biased and unbiased contexts, which ignores the gap between such evaluations and real-world use cases where users interact with LLMs by reading model responses and expect fair and safe outputs rather than LLMs' probabilities. To enable consistent evaluation across debiasing methods and bridge this gap, we introduce BiasFreeBench, an empirical benchmark that comprehensively compares eight mainstream bias mitigation techniques (covering four prompting-based and four training-based methods) on two test scenarios (multi-choice QA and open-ended multi-turn QA) by reorganizing existing datasets into a unified query-response setting. We further introduce a response-level metric, Bias-Free Score, to measure the extent to which LLM responses are fair, safe, and anti-stereotypical. Debiasing performances are systematically compared and analyzed across key dimensions: the prompting vs. training paradigm, model size, and generalization of different training strategies to unseen bias types. We will publicly release our benchmark, aiming to establish a unified testbed for bias mitigation research.
Authors: Monishwaran Maheswaran, Marco Carini, Christian Federmann, Tony Diaz
Abstract: We introduce TASER (Translation Assessment via Systematic Evaluation and Reasoning), a metric that uses Large Reasoning Models (LRMs) for automated translation quality assessment. TASER harnesses the explicit reasoning capabilities of LRMs to conduct systematic, step-by-step evaluation of translation quality. We evaluate TASER on the WMT24 Metrics Shared Task across both reference-based and reference-free scenarios, demonstrating state-of-the-art performance. In system-level evaluation, TASER achieves the highest soft pairwise accuracy in both reference-based and reference-free settings, outperforming all existing metrics. At the segment level, TASER maintains competitive performance with our reference-free variant ranking as the top-performing metric among all reference-free approaches. Our experiments reveal that structured prompting templates yield superior results with LRMs compared to the open-ended approaches that proved optimal for traditional LLMs. We evaluate o3, a large reasoning model from OpenAI, with varying reasoning efforts, providing insights into the relationship between reasoning depth and evaluation quality. The explicit reasoning process in LRMs offers interpretability and visibility, addressing a key limitation of existing automated metrics. Our results demonstrate that Large Reasoning Models show a measurable advancement in translation quality assessment, combining improved accuracy with transparent evaluation across diverse language pairs.
Authors: Xiaoyu Song, William Han, Tony Chen, Chaojing Duan, Michael A. Rosenberg, Emerson Liu, Ding Zhao
Abstract: Interest in generative Electrocardiogram-Language Models (ELMs) is growing, as they can produce textual responses conditioned on ECG signals and textual queries. Unlike traditional classifiers that output label probabilities, ELMs are more versatile, supporting domain-specific tasks (e.g., waveform analysis, diagnosis, prognosis) as well as general tasks (e.g., open-ended questions, dialogue). Retrieval-Augmented Generation (RAG), widely used in Large Language Models (LLMs) to ground LLM outputs in retrieved knowledge, helps reduce hallucinations and improve natural language generation (NLG). However, despite its promise, no open-source implementation or systematic study of RAG pipeline design for ELMs currently exists. To address this gap, we present the first open-source RAG pipeline for ELMs, along with baselines and ablation studies for NLG. Experiments on three public datasets show that ELMs with RAG consistently improves performance over non-RAG baselines and highlights key ELM design considerations. Our code is available at: https://github.com/willxxy/ECG-Bench.
Authors: Zhuohang Li, Xiaowei Li, Chengyu Huang, Guowang Li, Katayoon Goshvadi, Bo Dai, Dale Schuurmans, Paul Zhou, Hamid Palangi, Yiwen Song, Palash Goyal, Murat Kantarcioglu, Bradley A. Malin, Yuan Xue
Abstract: The alignment of large language models (LLMs) with human values increasingly relies on using other LLMs as automated judges, or ``autoraters''. However, their reliability is limited by a foundational issue: they are trained on discrete preference labels, forcing a single ground truth onto tasks that are often subjective, ambiguous, or nuanced. We argue that a reliable autorater must learn to model the full distribution of preferences defined by a target population. In this paper, we propose a general framework for calibrating probabilistic autoraters to any given preference distribution. We formalize the problem and present two learning methods tailored to different data conditions: 1) a direct supervised fine-tuning for dense, probabilistic labels, and 2) a reinforcement learning approach for sparse, binary labels. Our empirical results show that finetuning autoraters with a distribution-matching objective leads to verbalized probability predictions that are better aligned with the target preference distribution, with improved calibration and significantly lower positional bias, all while preserving performance on objective tasks.
Authors: Zhexiong Liu, Diane Litman
Abstract: Large Language Models (LLMs) have shown extraordinary success across various text generation tasks; however, their potential for simple yet essential text classification remains underexplored, as LLM pre-training tends to emphasize generation over classification. While LLMs with instruction tuning can transform classification into a generation task, they often struggle to categorize nuanced texts. One such example is text revision, which involves nuanced edits between pairs of texts. Although simply fine-tuning LLMs for revision classification seems plausible, it requires a large amount of revision annotations, which are exceptionally expensive and scarce in the community. To address this issue, we introduce a plug-and-play layer-wise parameter-efficient fine-tuning (PEFT) framework, i.e., IR-Tuning, which fine-tunes a subset of important LLM layers that are dynamically selected based on their gradient norm distribution, while freezing those of redundant layers. Extensive experiments suggest that IR-Tuning surpasses several layer-wise PEFT baselines over diverse text revisions, while achieving fast convergence, low GPU memory consumption, and effectiveness on small revision corpora.
Authors: Joe Barrow, Raj Patel, Misha Kharkovski, Ben Davies, Ryan Schmitt
Abstract: Black box large language models (LLMs) make information extraction (IE) easy to configure, but hard to trust. Unlike traditional information extraction pipelines, the information "extracted" is not guaranteed to be grounded in the document. To prevent this, this paper introduces the notion of a "safe passage": context generated by the LLM that is both grounded in the document and consistent with the extracted information. This is operationalized via a three-step pipeline, SafePassage, which consists of: (1) an LLM extractor that generates structured entities and their contexts from a document, (2) a string-based global aligner, and (3) a scoring model. Results show that using these three parts in conjunction reduces hallucinations by up to 85% on information extraction tasks with minimal risk of flagging non-hallucinations. High agreement between the SafePassage pipeline and human judgments of extraction quality mean that the pipeline can be dually used to evaluate LLMs. Surprisingly, results also show that using a transformer encoder fine-tuned on a small number of task-specific examples can outperform an LLM scoring model at flagging unsafe passages. These annotations can be collected in as little as 1-2 hours.
Authors: Ruochen Li, Jun Li, Bailiang Jian, Kun Yuan, Youxiang Zhu
Abstract: Automatically generated radiology reports often receive high scores from existing evaluation metrics but fail to earn clinicians' trust. This gap reveals fundamental flaws in how current metrics assess the quality of generated reports. We rethink the design and evaluation of these metrics and propose a clinically grounded Meta-Evaluation framework. We define clinically grounded criteria spanning clinical alignment and key metric capabilities, including discrimination, robustness, and monotonicity. Using a fine-grained dataset of ground truth and rewritten report pairs annotated with error types, clinical significance labels, and explanations, we systematically evaluate existing metrics and reveal their limitations in interpreting clinical semantics, such as failing to distinguish clinically significant errors, over-penalizing harmless variations, and lacking consistency across error severity levels. Our framework offers guidance for building more clinically reliable evaluation methods.
Authors: \v{L}ubo\v{s} Kri\v{s}, Jaroslav Kop\v{c}an, Qiwei Peng, Andrej Ridzik, Marcel Vesel\'y, Martin Tamajka
Abstract: The proliferation of transformer-based language models has revolutionized NLP domain while simultaneously introduced significant challenges regarding model transparency and trustworthiness. The complexity of achieving explainable systems in this domain is evidenced by the extensive array of explanation methods and evaluation metrics developed by researchers. To address the challenge of selecting optimal explainability approaches, we present \textbf{\texttt{o-mega}}, a hyperparameter optimization tool designed to automatically identify the most effective explainable AI methods and their configurations within the semantic matching domain. We evaluate o-mega on a post-claim matching pipeline using a curated dataset of social media posts paired with refuting claims. Our tool systematically explores different explainable methods and their hyperparameters, demonstrating improved transparency in automated fact-checking systems. As a result, such automated optimization of explanation methods can significantly enhance the interpretability of claim-matching models in critical applications such as misinformation detection, contributing to more trustworthy and transparent AI systems.
Authors: Bowen Wei, Yuan Shen Tay, Howard Liu, Jinhao Pan, Kun Luo, Ziwei Zhu, Chris Jordan
Abstract: Security Operations Centers (SOCs) are overwhelmed by tens of thousands of daily alerts, with only a small fraction corresponding to genuine attacks. This overload creates alert fatigue, leading to overlooked threats and analyst burnout. Classical detection pipelines are brittle and context-poor, while recent LLM-based approaches typically rely on a single model to interpret logs, retrieve context, and adjudicate alerts end-to-end -- an approach that struggles with noisy enterprise data and offers limited transparency. We propose CORTEX, a multi-agent LLM architecture for high-stakes alert triage in which specialized agents collaborate over real evidence: a behavior-analysis agent inspects activity sequences, evidence-gathering agents query external systems, and a reasoning agent synthesizes findings into an auditable decision. To support training and evaluation, we release a dataset of fine-grained SOC investigations from production environments, capturing step-by-step analyst actions and linked tool outputs. Across diverse enterprise scenarios, CORTEX substantially reduces false positives and improves investigation quality over state-of-the-art single-agent LLMs.
Authors: Zijun Wu, Yongchang Hao, Lili Mou
Abstract: Large language models rely heavily on prompts to specify tasks, recall knowledge and guide reasoning. However, this reliance is inefficient as prompts must be re-read at each step, scale poorly across tasks, and lack mechanisms for modular reuse. We introduce TokMem, a tokenized procedural memory that stores recurring procedures as compact, trainable embeddings. Each memory token encodes both an address to a procedure and a control signal that steers generation, enabling targeted behavior with constant-size overhead. To support continual adaptation, TokMem keeps the backbone model frozen, allowing new procedures to be added without interfering with existing ones. We evaluate TokMem on 1,000 tasks for atomic recall, and on function-calling tasks for compositional recall, where it consistently outperforms retrieval-augmented generation while avoiding repeated context overhead, and fine-tuning with far fewer parameters. These results establish TokMem as a scalable and modular alternative to prompt engineering and fine-tuning, offering an explicit procedural memory for LLMs.
Authors: Yuling Shi, Yichun Qian, Hongyu Zhang, Beijun Shen, Xiaodong Gu
Abstract: Code generation under long contexts is becoming increasingly critical as Large Language Models (LLMs) are required to reason over extensive information in the codebase. While recent advances enable code LLMs to process long inputs, high API costs and generation latency remain substantial bottlenecks. Existing context pruning techniques, such as LLMLingua, achieve promising results for general text but overlook code-specific structures and dependencies, leading to suboptimal performance in programming tasks. In this paper, we propose LongCodeZip, a novel plug-and-play code compression framework designed specifically for code LLMs. LongCodeZip employs a dual-stage strategy: (1) coarse-grained compression, which identifies and ranks function-level chunks using conditional perplexity with respect to the instruction, retaining only the most relevant functions; and (2) fine-grained compression, which segments retained functions into blocks based on perplexity and selects an optimal subset under an adaptive token budget to maximize relevance. Evaluations across multiple tasks, including code completion, summarization, and question answering, show that LongCodeZip consistently outperforms baseline methods, achieving up to a 5.6x compression ratio without degrading task performance. By effectively reducing context size while preserving essential information, LongCodeZip enables LLMs to better scale to real-world, large-scale code scenarios, advancing the efficiency and capability of code intelligence applications.
Authors: Koki Ryu, Hitomi Yanaka
Abstract: Personalizing the outputs of large language models (LLMs) to align with individual user preferences is an active research area. However, previous studies have mainly focused on classification or ranking tasks and have not considered Likert-scale rating prediction, a regression task that requires both language and mathematical reasoning to be solved effectively. This task has significant industrial applications, but the utilization of LLMs remains underexplored, particularly regarding the capabilities of off-the-shelf LLMs. This study investigates the performance of off-the-shelf LLMs on rating prediction, providing different in-context information. Through comprehensive experiments with eight models across three datasets, we demonstrate that user-written reviews significantly improve the rating prediction performance of LLMs. This result is comparable to traditional methods like matrix factorization, highlighting the potential of LLMs as a promising solution for the cold-start problem. We also find that the reviews for concrete items are more effective than general preference descriptions that are not based on any specific item. Furthermore, we discover that prompting LLMs to first generate a hypothetical review enhances the rating prediction performance. Our code is available at https://github.com/ynklab/rating-prediction-with-reviews.
URLs: https://github.com/ynklab/rating-prediction-with-reviews.
Authors: Yawen Xue, Masaya Tsunokake, Yuta Koreeda, Ekant Muljibhai Amin, Takashi Sumiyoshi, Yasuhiro Sogawa
Abstract: Agentic large language models (LLMs) have become prominent for autonomously interacting with external environments and performing multi-step reasoning tasks. Most approaches leverage these capabilities via in-context learning with few-shot prompts, but this often results in lengthy inputs and higher computational costs. Agent fine-tuning offers an alternative by enabling LLMs to internalize procedural reasoning and domain-specific knowledge through training on relevant data and demonstration trajectories. While prior studies have focused on general domains, their effectiveness in specialized technical microdomains remains unclear. This paper explores agent fine-tuning for domain adaptation within Hitachi's JP1 middleware, a microdomain for specialized IT operations. We fine-tuned LLMs using JP1-specific datasets derived from domain manuals and distilled reasoning trajectories generated by LLMs themselves, enhancing decision making accuracy and search efficiency. During inference, we used an agentic prompt with retrieval-augmented generation and introduced a context-answer extractor to improve information relevance. On JP1 certification exam questions, our method achieved a 14% performance improvement over the base model, demonstrating the potential of agent fine-tuning for domain-specific reasoning in complex microdomains.
Authors: Pengzhou Cheng, Lingzhong Dong, Zeng Wu, Zongru Wu, Xiangru Tang, Chengwei Qin, Zhuosheng Zhang, Gongshen Liu
Abstract: Although numerous strategies have recently been proposed to enhance the autonomous interaction capabilities of multimodal agents in graphical user interface (GUI), their reliability remains limited when faced with complex or out-of-domain tasks. This raises a fundamental question: Are existing multimodal agents reasoning spuriously? In this paper, we propose \textbf{Agent-ScanKit}, a systematic probing framework to unravel the memory and reasoning capabilities of multimodal agents under controlled perturbations. Specifically, we introduce three orthogonal probing paradigms: visual-guided, text-guided, and structure-guided, each designed to quantify the contributions of memorization and reasoning without requiring access to model internals. In five publicly available GUI benchmarks involving 18 multimodal agents, the results demonstrate that mechanical memorization often outweighs systematic reasoning. Most of the models function predominantly as retrievers of training-aligned knowledge, exhibiting limited generalization. Our findings underscore the necessity of robust reasoning modeling for multimodal agents in real-world scenarios, offering valuable insights toward the development of reliable multimodal agents.
Authors: Xingjian Zhao, Zhe Xu, Luozhijie Jin, Yang Wang, Hanfu Chen, Yaozhou Jiang, Ke Chen, Ruixiao Li, Mingshu Chen, Ruiming Wang, Wenbo Zhang, Yiyang Zhang, Donghua Yu, Yang Gao, Xiaogui Yang, Yitian Gong, Yuanfan Xu, Qinyuan Cheng, Zhaoye Fei, Shimin Li, Yaqian Zhou, Xuanjing Huang, Xipeng Qiu
Abstract: Spoken dialogue systems often rely on cascaded pipelines that transcribe, process, and resynthesize speech. While effective, this design discards paralinguistic cues and limits expressivity. Recent end-to-end methods reduce latency and better preserve these cues, yet still rely on text intermediates, creating a fundamental bottleneck. We present MOSS-Speech, a true speech-to-speech large language model that directly understands and generates speech without relying on text guidance. Our approach combines a modality-based layer-splitting architecture with a frozen pre-training strategy, preserving the reasoning and knowledge of pretrained text LLMs while adding native speech capabilities. Experiments show that our model achieves state-of-the-art results in spoken question answering and delivers comparable speech-to-speech performance relative to existing text-guided systems, while still maintaining competitive text performance. By narrowing the gap between text-guided and direct speech generation, our work establishes a new paradigm for expressive and efficient end-to-end speech interaction.
Authors: Yurun Chen, Xavier Hu, Yuhan Liu, Ziqi Wang, Zeyi Liao, Lin Chen, Feng Wei, Yuxi Qian, Bo Zheng, Keting Yin, Shengyu Zhang
Abstract: As multimodal LLM-driven agents continue to advance in autonomy and generalization, evaluation based on static datasets can no longer adequately assess their true capabilities in dynamic environments and diverse tasks. Existing LLM-based synthetic data methods are largely designed for LLM training and evaluation, and thus cannot be directly applied to agent tasks that require tool use and interactive capabilities. While recent studies have explored automatic agent task generation with LLMs, most efforts remain limited to text or image analysis, without systematically modeling multi-step interactions in web environments. To address these challenges, we propose Graph2Eval, a knowledge graph-based framework that automatically generates both multimodal document comprehension tasks and web interaction tasks, enabling comprehensive evaluation of agents' reasoning, collaboration, and interactive capabilities. In our approach, knowledge graphs constructed from multi-source external data serve as the task space, where we translate semantic relations into structured multimodal tasks using subgraph sampling, task templates, and meta-paths. A multi-stage filtering pipeline based on node reachability, LLM scoring, and similarity analysis is applied to guarantee the quality and executability of the generated tasks. Furthermore, Graph2Eval supports end-to-end evaluation of multiple agent types (Single-Agent, Multi-Agent, Web Agent) and measures reasoning, collaboration, and interaction capabilities. We instantiate the framework with Graph2Eval-Bench, a curated dataset of 1,319 tasks spanning document comprehension and web interaction scenarios. Experiments show that Graph2Eval efficiently generates tasks that differentiate agent and model performance, revealing gaps in reasoning, collaboration, and web interaction across different settings and offering a new perspective for agent evaluation.
Authors: Yongchao Long, Xian Wu, Yingying Zhang, Xianbin Wen, Yuxi Zhou, Shenda Hong
Abstract: While Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to generate contextually grounded responses, contextual faithfulness remains challenging as LLMs may not consistently trust provided context, leading to hallucinations that undermine reliability. We observe an inverse correlation between response copying degree and context-unfaithful hallucinations on RAGTruth, suggesting that higher copying degrees reduce hallucinations by fostering genuine contextual belief. We propose CopyPasteLLM, obtained through two-stage high-copying response preference training. We design three prompting methods to enhance copying degree, demonstrating that high-copying responses achieve superior contextual faithfulness and hallucination control. These approaches enable a fully automated pipeline that transforms generated responses into high-copying preference data for training CopyPasteLLM. On FaithEval, ConFiQA and PubMedQA, CopyPasteLLM achieves best performance in both counterfactual and original contexts, remarkably with 12.2% to 24.5% accuracy improvements on FaithEval over the best baseline, while requiring only 365 training samples -- 1/50th of baseline data. To elucidate CopyPasteLLM's effectiveness, we propose the Context-Parameter Copying Capturing algorithm. Interestingly, this reveals that CopyPasteLLM recalibrates reliance on internal parametric knowledge rather than external knowledge during generation. All codes are available at https://github.com/longyongchao/CopyPasteLLM
Authors: Jiarun Liu, Shiyue Xu, Shangkun Liu, Yang Li, Wen Liu, Min Liu, Xiaoqing Zhou, Hanmin Wang, Shilin Jia, zhen Wang, Shaohua Tian, Hanhao Li, Junbo Zhang, Yongli Yu, Peng Cao, Haofen Wang
Abstract: Large Language Models are increasingly deployed as autonomous agents for complex real-world tasks, yet existing systems often focus on isolated improvements without a unifying design for robustness and adaptability. We propose a generalist agent architecture that integrates three core components: a collective multi-agent framework combining planning and execution agents with critic model voting, a hierarchical memory system spanning working, semantic, and procedural layers, and a refined tool suite for search, code execution, and multimodal parsing. Evaluated on a comprehensive benchmark, our framework consistently outperforms open-source baselines and approaches the performance of proprietary systems. These results demonstrate the importance of system-level integration and highlight a path toward scalable, resilient, and adaptive AI assistants capable of operating across diverse domains and tasks.
Authors: Samuel Pfisterer, Florian Gr\"otschla, Luca A. Lanzend\"orfer, Florian Yan, Roger Wattenhofer
Abstract: Recent progress in speech processing has highlighted that high-quality performance across languages requires substantial training data for each individual language. While existing multilingual datasets cover many languages, they often contain insufficient data for most languages. Thus, trained models perform poorly on the majority of the supported languages. Our work addresses this challenge by introducing a scalable pipeline for constructing speech datasets from parliamentary recordings. The proposed pipeline includes robust components for media retrieval and a two-stage alignment algorithm designed to handle non-verbatim transcripts and long-form audio. Applying this pipeline to recordings from 22 European parliaments, we extract over 61k hours of aligned speech segments, achieving substantial per-language coverage with 19 languages exceeding 1k hours and 22 languages exceeding 500 hours of high-quality speech data. We obtain an average 41.8\% reduction in word error rates over baselines when finetuning an existing ASR model on our dataset, demonstrating the usefulness of our approach.
Authors: Gaotang Li, Ruizhong Qiu, Xiusi Chen, Heng Ji, Hanghang Tong
Abstract: Supervised fine-tuning (SFT) is the standard approach for post-training large language models (LLMs), yet it often shows limited generalization. We trace this limitation to its default training objective: negative log likelihood (NLL). While NLL is classically optimal when training from scratch, post-training operates in a different paradigm and could violate its optimality assumptions, where models already encode task-relevant priors and supervision can be long and noisy. To this end, we study a general family of probability-based objectives and characterize their effectiveness under different conditions. Through comprehensive experiments and extensive ablation studies across 7 model backbones, 14 benchmarks, and 3 domains, we uncover a critical dimension that governs objective behavior: the model-capability continuum. Near the model-strong end, prior-leaning objectives that downweight low-probability tokens (e.g., $-p$, $-p^{10}$, thresholded variants) consistently outperform NLL; toward the model-weak end, NLL dominates; in between, no single objective prevails. Our theoretical analysis further elucidates how objectives trade places across the continuum, providing a principled foundation for adapting objectives to model capability. Our code is available at https://github.com/GaotangLi/Beyond-Log-Likelihood.
Authors: Kung-Hsiang Huang, Haoyi Qiu, Yutong Dai, Caiming Xiong, Chien-Sheng Wu
Abstract: Graphical user interface (GUI) agents built on vision-language models have emerged as a promising approach to automate human-computer workflows. However, they also face the inefficiency challenge as they process long sequences of high-resolution screenshots and solving long-horizon tasks, making inference slow, costly and memory-bound. While key-value (KV) caching can mitigate this, storing the full cache is prohibitive for image-heavy contexts. Existing cache-compression methods are sub-optimal as they do not account for the spatial and temporal redundancy of GUIs. In this work, we first analyze attention patterns in GUI agent workloads and find that, unlike in natural images, attention sparsity is uniformly high across all transformer layers. This insight motivates a simple uniform budget allocation strategy, which we show empirically outperforms more complex layer-varying schemes. Building on this, we introduce GUI-KV, a plug-and-play KV cache compression method for GUI agents that requires no retraining. GUI-KV combines two novel techniques: (i) spatial saliency guidance, which augments attention scores with the L2 norm of hidden states to better preserve semantically important visual tokens, and (ii) temporal redundancy scoring, which projects previous frames' keys onto the current frame's key subspace to preferentially prune redundant history. Across standard GUI agent benchmarks and models, GUI-KV outperforms competitive KV compression baselines, closely matching full-cache accuracy at modest budgets. Notably, in a 5-screenshot setting on the AgentNetBench benchmark, GUI-KV reduces decoding FLOPs by 38.9% while increasing step accuracy by 4.1% over the full-cache baseline. These results demonstrate that exploiting GUI-specific redundancies enables efficient and reliable agent performance.
Authors: Minjae Oh, Sangjun Song, Seungkyu Lee, Sungmin Jo, Yohan Jo
Abstract: Small reasoning models (SRMs) often overthink during tool use: they reach a correct tool-argument configuration, then continue reasoning and overwrite it with an incorrect final call. We diagnose overthinking via oracle rollouts that inject at sentence boundaries. On the Berkeley Function Calling Leaderboard (BFCL), this oracle termination lifts average accuracy from 85.8\% to 94.2\% while reducing tokens by 80-94\%, revealing substantial recoverable headroom and potential redundant reasoning. While prior work on concise reasoning has largely targeted mathematics, tool reasoning remains underexplored. We adapt various early-termination baselines to tool use and introduce ThinkBrake, a training-free decoding heuristic. ThinkBrake monitors the log-probability margin between and the current top token at sentence boundaries and triggers termination when this margin becomes small. Across BFCL's single turn, non-live and live splits, ThinkBrake preserves or improves accuracy while reducing tokens up to 25\%, outperforming various baselines.
Authors: Yubo Xie, Chenkai Wang, Zongyang Ma, Fahui Miao
Abstract: Large language models (LLMs) are trained on vast amounts of text from the Internet, but do they truly understand the viral content that rapidly spreads online -- commonly known as memes? In this paper, we introduce CHIME, a dataset for CHinese Internet Meme Explanation. The dataset comprises popular phrase-based memes from the Chinese Internet, annotated with detailed information on their meaning, origin, example sentences, types, etc. To evaluate whether LLMs understand these memes, we designed two tasks. In the first task, we assessed the models' ability to explain a given meme, identify its origin, and generate appropriate example sentences. The results show that while LLMs can explain the meanings of some memes, their performance declines significantly for culturally and linguistically nuanced meme types. Additionally, they consistently struggle to provide accurate origins for the memes. In the second task, we created a set of multiple-choice questions (MCQs) requiring LLMs to select the most appropriate meme to fill in a blank within a contextual sentence. While the evaluated models were able to provide correct answers, their performance remains noticeably below human levels. We have made CHIME public and hope it will facilitate future research on computational meme understanding.
Authors: Shiyu Li, Yang Tang, Yifan Wang, Peiming Li, Xi Chen
Abstract: Search agents powered by Large Language Models (LLMs) have demonstrated significant potential in tackling knowledge-intensive tasks. Reinforcement learning (RL) has emerged as a powerful paradigm for training these agents to perform complex, multi-step reasoning. However, prior RL-based methods often rely on sparse or rule-based rewards, which can lead agents to commit to suboptimal or erroneous reasoning paths without the ability to recover. To address these limitations, we propose ReSeek, a novel self-correcting framework for training search agents. Our framework introduces a self-correction mechanism that empowers the agent to dynamically identify and recover from erroneous search paths during an episode. By invoking a special JUDGE action, the agent can judge the information and re-plan its search strategy. To guide this process, we design a dense, instructive process reward function, which decomposes into a correctness reward for retrieving factual information and a utility reward for finding information genuinely useful for the query. Furthermore, to mitigate the risk of data contamination in existing datasets, we introduce FictionalHot, a new and challenging benchmark with recently curated questions requiring complex reasoning. Being intuitively reasonable and practically simple, extensive experiments show that agents trained with ReSeek significantly outperform SOTA baselines in task success rate and path faithfulness.
Authors: Li Li, Ziyi Wang, Yongliang Wu, Jianfei Cai, Xu Yang
Abstract: Chain-of-Thought (CoT) prompting has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs). However, existing implementations, such as in-context learning and fine-tuning, remain costly and inefficient. To improve CoT reasoning at a lower cost, and inspired by the task vector paradigm, we introduce CoT Vectors, compact representations that encode task-general, multi-step reasoning knowledge. Through experiments with Extracted CoT Vectors, we observe pronounced layer-wise instability, manifesting as a U-shaped performance curve that reflects a systematic three-stage reasoning process in LLMs. To address this limitation, we propose Learnable CoT Vectors, optimized under a teacher-student framework to provide more stable and robust guidance. Extensive evaluations across diverse benchmarks and models demonstrate that CoT Vectors not only outperform existing baselines but also achieve performance comparable to parameter-efficient fine-tuning methods, while requiring fewer trainable parameters. Moreover, by treating CoT Vectors as a probe, we uncover how their effectiveness varies due to latent space structure, information density, acquisition mechanisms, and pre-training differences, offering new insights into the functional organization of multi-step reasoning in LLMs. The source code will be released.
Authors: Sangmin Lee, Woongjib Choi, Jihyun Kim, Hong-Goo Kang
Abstract: In this paper, we present a neural spoken language diarization model that supports an unconstrained span of languages within a single framework. Our approach integrates a learnable query-based architecture grounded in multilingual awareness, with large-scale pretraining on simulated code-switching data. By jointly leveraging these two components, our method overcomes the limitations of conventional approaches in data scarcity and architecture optimization, and generalizes effectively to real-world multilingual settings across diverse environments. Experimental results demonstrate that our approach achieves state-of-the-art performance on several language diarization benchmarks, with a relative performance improvement of 23% to 52% over previous methods. We believe that this work not only advances research in language diarization but also establishes a foundational framework for code-switching speech technologies.
Authors: Teisovi Angami, Kevisino Khate
Abstract: The Tenyidie language is a low-resource language of the Tibeto-Burman family spoken by the Tenyimia Community of Nagaland in the north-eastern part of India and is considered a major language in Nagaland. It is tonal, Subject-Object-Verb, and highly agglutinative in nature. Being a low-resource language, very limited research on Natural Language Processing (NLP) has been conducted. To the best of our knowledge, no work on syllabification has been reported for this language. Among the many NLP tasks, syllabification or syllabication is an important task in which the given word syllables are identified. The contribution of this work is the creation of 10,120 syllabified Tenyidie words and the application of the Deep Learning techniques on the created corpus. In this paper, we have applied LSTM, BLSTM, BLSTM+CRF, and Encoder-decoder deep learning architectures on our created dataset. In our dataset split of 80:10:10 (train:validation:test) set, we achieved the highest accuracy of 99.21% with BLSTM model on the test set. This work will find its application in numerous other NLP applications, such as morphological analysis, part-of-speech tagging, machine translation, etc, for the Tenyidie Language. Keywords: Tenyidie; NLP; syllabification; deep learning; LSTM; BLSTM; CRF; Encoder-decoder
Authors: Jinlan Fu, Shenzhen Huangfu, Hao Fei, Yichong Huang, Xiaoyu Shen, Xipeng Qiu, See-Kiong Ng
Abstract: The alt-text generation task produces concise, context-relevant descriptions of images, enabling blind and low-vision users to access online images. Despite the capabilities of large vision-language models, alt-text generation performance remains limited due to noisy user annotations, inconsistent standards, and MLLMs' insensitivity to contextual information. Previous efforts to fine-tune MLLMs using supervised fine-tuning (SFT) have struggled, as SFT relies on accurate target annotations, which are often flawed in user-generated alt-text. To address this, we propose Multi-faceted Cross-modal Direct Preference Optimization (MCM-DPO), which improves alt-text generation by learning to identify better options in preference pairs without requiring precise annotations. MCM-DPO optimizes preferences across single, paired, and multi-preference dimensions, covering textual, visual, and cross-modal factors. In light of the scarcity of high-quality annotated and preference-labeled datasets for alt-text, we constructed two large-scale, high-quality datasets named TAlt and PAlt, sourced from Twitter and Pinterest. These datasets include 202k annotated alt-text samples and 18k preference pairs that cover diverse preference dimensions, aiming to support further research in this domain. Experimental results show that our proposed MCM-DPO method consistently outperforms both DPO and SFT, establishing a new state of the art in alt-text generation. We release the code and data here: https://github.com/LVUGAI/MCM-DPO
Authors: Fran\c{c}ois Ledoyen, Ga\"el Dias, Jeremie Pantin, Alexis Lechervy, Fabrice Maurel, Youssef Chahir
Abstract: Simplifying complex texts is essential for ensuring equitable access to information, especially for individuals with cognitive impairments. The Easy-to-Read (ETR) initiative offers a framework for making content accessible to the neurodivergent population, but the manual creation of such texts remains time-consuming and resource-intensive. In this work, we investigate the potential of large language models (LLMs) to automate the generation of ETR content. To address the scarcity of aligned corpora and the specificity of ETR constraints, we propose a multi-task learning (MTL) approach that trains models jointly on text summarization, text simplification, and ETR generation. We explore two different strategies: multi-task retrieval-augmented generation (RAG) for in-context learning, and MTL-LoRA for parameter-efficient fine-tuning. Our experiments with Mistral-7B and LLaMA-3-8B, based on ETR-fr, a new high-quality dataset, demonstrate the benefits of multi-task setups over single-task baselines across all configurations. Moreover, results show that the RAG-based strategy enables generalization in out-of-domain settings, while MTL-LoRA outperforms all learning strategies within in-domain configurations.
Authors: Fran\c{c}ois Ledoyen, Ga\"el Dias, Alexis Lechervy, Jeremie Pantin, Fabrice Maurel, Youssef Chahir, Elisa Gouzonnat, M\'elanie Berthelot, Stanislas Moravac, Armony Altinier, Amy Khairalla
Abstract: Ensuring accessibility for individuals with cognitive impairments is essential for autonomy, self-determination, and full citizenship. However, manual Easy-to-Read (ETR) text adaptations are slow, costly, and difficult to scale, limiting access to crucial information in healthcare, education, and civic life. AI-driven ETR generation offers a scalable solution but faces key challenges, including dataset scarcity, domain adaptation, and balancing lightweight learning of Large Language Models (LLMs). In this paper, we introduce ETR-fr, the first dataset for ETR text generation fully compliant with European ETR guidelines. We implement parameter-efficient fine-tuning on PLMs and LLMs to establish generative baselines. To ensure high-quality and accessible outputs, we introduce an evaluation framework based on automatic metrics supplemented by human assessments. The latter is conducted using a 36-question evaluation form that is aligned with the guidelines. Overall results show that PLMs perform comparably to LLMs and adapt effectively to out-of-domain texts.
Authors: Harethah Abu Shairah, Somayah AlHarbi, Abdulaziz AlHussein, Sameer Alsabea, Omar Shaqaqi, Hebah AlShamlan, Omar Knio, George Turkiyyah
Abstract: We introduce ALARB, a dataset and suite of tasks designed to evaluate the reasoning capabilities of large language models (LLMs) within the Arabic legal domain. While existing Arabic benchmarks cover some knowledge-intensive tasks such as retrieval and understanding, substantial datasets focusing specifically on multistep reasoning for Arabic LLMs, especially in open-ended contexts, are lacking. The dataset comprises over 13K commercial court cases from Saudi Arabia, with each case including the facts presented, the reasoning of the court, the verdict, as well as the cited clauses extracted from the regulatory documents. We define a set of challenging tasks leveraging this dataset and reflecting the complexity of real-world legal reasoning, including verdict prediction, completion of reasoning chains in multistep legal arguments, and identification of relevant regulations based on case facts. We benchmark a representative selection of current open and closed Arabic LLMs on these tasks and demonstrate the dataset's utility for instruction tuning. Notably, we show that instruction-tuning a modest 12B parameter model using ALARB significantly enhances its performance in verdict prediction and Arabic verdict generation, reaching a level comparable to that of GPT-4o.
Authors: Jenny Kunz, Iben Nyholm Debess, Annika Simonsen
Abstract: We investigate how to adapt small, efficient LLMs to Faroese, a low-resource North Germanic language. Starting from English models, we continue pre-training on related Scandinavian languages, either individually or combined via merging, before fine-tuning on Faroese. We compare full fine-tuning with parameter-efficient tuning using LoRA, evaluating their impact on both linguistic accuracy and text comprehension. Due to the lack of existing Faroese evaluation data, we construct two new minimal-pair benchmarks from adapted and newly collected datasets and complement them with human evaluations by Faroese linguists. Our results demonstrate that transfer from related languages is crucial, though the optimal source language depends on the task: Icelandic enhances linguistic accuracy, whereas Danish boosts comprehension. Similarly, the choice between full fine-tuning and LoRA is task-dependent: LoRA improves linguistic acceptability and slightly increases human evaluation scores on the base model, while full fine-tuning yields stronger comprehension performance and better preserves model capabilities during downstream fine-tuning.
Authors: Yanming Sun, Runzhe Zhan, Chi Seng Cheang, Han Wu, Xuebo Liu, Yuyao Niu, Fengying Ye, Kaixin Lan, Lidia S. Chao, Derek F. Wong
Abstract: \textbf{RE}trieval-\textbf{A}ugmented \textbf{L}LM-based \textbf{M}achine \textbf{T}ranslation (REAL-MT) shows promise for knowledge-intensive tasks like idiomatic translation, but its reliability under noisy retrieval contexts remains poorly understood despite this being a common challenge in real-world deployment. To address this gap, we propose a noise synthesis framework and new metrics to evaluate the robustness of REAL-MT systematically. Using this framework, we instantiate REAL-MT with Qwen-series models, including standard LLMs and large reasoning models (LRMs) with enhanced reasoning, and evaluate their performance on idiomatic translation across high-, medium-, and low-resource language pairs under synthesized noise. Our results show that low-resource language pairs, which rely more heavily on retrieved context, degrade more severely under noise than high-resource ones and often produce nonsensical translations. Although LRMs possess enhanced reasoning capabilities, they show no improvement in error correction and are even more susceptible to noise, tending to rationalize incorrect contexts. We find that this stems from an attention shift away from the source idiom to noisy content, while confidence increases despite declining accuracy, indicating poor calibration. To mitigate these issues, we investigate training-free and fine-tuning strategies, which improve robustness at the cost of performance in clean contexts, revealing a fundamental trade-off. Our findings highlight the limitations of current approaches, underscoring the need for self-verifying integration mechanisms.
Authors: Adi Simhi, Jonathan Herzig, Martin Tutek, Itay Itzhak, Idan Szpektor, Yonatan Belinkov
Abstract: As large language models (LLMs) evolve from conversational assistants into autonomous agents, evaluating the safety of their actions becomes critical. Prior safety benchmarks have primarily focused on preventing generation of harmful content, such as toxic text. However, they overlook the challenge of agents taking harmful actions when the most effective path to an operational goal conflicts with human safety. To address this gap, we introduce ManagerBench, a benchmark that evaluates LLM decision-making in realistic, human-validated managerial scenarios. Each scenario forces a choice between a pragmatic but harmful action that achieves an operational goal, and a safe action that leads to worse operational performance. A parallel control set, where potential harm is directed only at inanimate objects, measures a model's pragmatism and identifies its tendency to be overly safe. Our findings indicate that the frontier LLMs perform poorly when navigating this safety-pragmatism trade-off. Many consistently choose harmful options to advance their operational goals, while others avoid harm only to become overly safe and ineffective. Critically, we find this misalignment does not stem from an inability to perceive harm, as models' harm assessments align with human judgments, but from flawed prioritization. ManagerBench is a challenging benchmark for a core component of agentic behavior: making safe choices when operational goals and alignment values incentivize conflicting actions. Benchmark & code available at https://github.com/technion-cs-nlp/ManagerBench.
Authors: Ziliang Wang, Kang An, Xuhui Zheng, Faqiang Qian, Weikun Zhang, Cijun Ouyang, Jialu Cai, Yuhang Wang, Yichao Wu
Abstract: While search-augmented large language models (LLMs) exhibit impressive capabilities, their reliability in complex multi-hop reasoning remains limited. This limitation arises from three fundamental challenges: decomposition errors, where tasks are incorrectly broken down; retrieval missing, where key evidence fails to be retrieved; and reasoning errors, where flawed logic propagates through the reasoning chain. A single failure in any of these stages can derail the final answer. We propose Erasable Reinforcement Learning (ERL), a novel framework that transforms fragile reasoning into a robust process. ERL explicitly identifies faulty steps, erases them, and regenerates reasoning in place, preventing defective logic from propagating through the reasoning chain. This targeted correction mechanism turns brittle reasoning into a more resilient process. Models trained with ERL, termed ESearch, achieve substantial improvements on HotpotQA, MuSiQue, 2Wiki, and Bamboogle, with the 3B model achieving +8.48% EM and +11.56% F1, and the 7B model achieving +5.38% EM and +7.22% F1 over previous state-of-the-art(SOTA) results. These findings suggest that erasable reinforcement learning provides a powerful paradigm shift for robust multi-step reasoning in LLMs.
Authors: Loris Bergeron, Ioana Buhnila, J\'er\^ome Fran\c{c}ois, Radu State
Abstract: Large Language Models (LLMs) excel in many NLP tasks but remain prone to hallucinations, limiting trust in real-world applications. We present HalluGuard, a 4B-parameter Small Reasoning Model (SRM) for mitigating hallucinations in Retrieval-Augmented Generation (RAG). HalluGuard classifies document-claim pairs as grounded or hallucinated and produces evidence-grounded justifications for transparency. Our approach combines (i) a domain-agnostic synthetic dataset derived from FineWeb and refined through multi-stage curation and data reformation, (ii) synthetic grounded and hallucinated claims, and (iii) preference-based fine-tuning with Odds Ratio Preference Optimization to distill large-model reasoning into a smaller backbone. On the RAGTruth subset of the LLM-AggreFact benchmark, HalluGuard achieves 84.0% balanced accuracy (BAcc), rivaling specialized models, MiniCheck (7B; 84.0%) and Granite Guardian 3.3 (8B; 82.2%) while using roughly half their parameters. Over the full benchmark it reaches 75.7% BAcc, matching larger general-purpose LLMs such as GPT-4o (75.9%). We will release HalluGuard and datasets under Apache 2.0 upon acceptance.
Authors: Zhen Yin, Shenghua Wang
Abstract: The rapid adoption of large language models (LLMs) in scientific writing raises serious concerns regarding authorship integrity and the reliability of scholarly publications. Existing detection approaches mainly rely on document-level classification or surface-level statistical cues; however, they neglect fine-grained span localization, exhibit weak calibration, and often fail to generalize across disciplines and generators. To address these limitations, we present Sci-SpanDet, a structure-aware framework for detecting AI-generated scholarly texts. The proposed method combines section-conditioned stylistic modeling with multi-level contrastive learning to capture nuanced human-AI differences while mitigating topic dependence, thereby enhancing cross-domain robustness. In addition, it integrates BIO-CRF sequence labeling with pointer-based boundary decoding and confidence calibration to enable precise span-level detection and reliable probability estimates. Extensive experiments on a newly constructed cross-disciplinary dataset of 100,000 annotated samples generated by multiple LLM families (GPT, Qwen, DeepSeek, LLaMA) demonstrate that Sci-SpanDet achieves state-of-the-art performance, with F1(AI) of 80.17, AUROC of 92.63, and Span-F1 of 74.36. Furthermore, it shows strong resilience under adversarial rewriting and maintains balanced accuracy across IMRaD sections and diverse disciplines, substantially surpassing existing baselines. To ensure reproducibility and to foster further research on AI-generated text detection in scholarly documents, the curated dataset and source code will be publicly released upon publication.
Authors: Shunfeng Zheng, Yudi Zhang, Meng Fang, Zihan Zhang, Zhitan Wu, Mykola Pechenizkiy, Ling Chen
Abstract: Retrieval-augmented generation (RAG) with foundation models has achieved strong performance across diverse tasks, but their capacity for expert-level reasoning-such as solving Olympiad-level physics problems-remains largely unexplored. Inspired by the way students prepare for competitions by reviewing past problems, we investigate the potential of RAG to enhance physics reasoning in foundation models. We introduce PhoPile, a high-quality multimodal dataset specifically designed for Olympiad-level physics, enabling systematic study of retrieval-based reasoning. PhoPile includes diagrams, graphs, and equations, capturing the inherently multimodal nature of physics problem solving. Using PhoPile, we benchmark RAG-augmented foundation models, covering both large language models (LLMs) and large multimodal models (LMMs) with multiple retrievers. Our results demonstrate that integrating retrieval with physics corpora can improve model performance, while also highlighting challenges that motivate further research in retrieval-augmented physics reasoning.
Authors: Ammar Khairi, Daniel D'souza, Marzieh Fadaee, Julia Kreutzer
Abstract: Obtaining high-quality generations in modern LLMs has largely been framed as a selection problem: identifying a single winning generation from a diverse pool of N samples, the Best-of-N (BoN). Yet, this approach is inherently zero-sum, discarding diverse and potentially useful information from the pool. Instead, we explore a collaborative setup, where all candidates can potentially contribute to the final winning generation. To this end, we propose Fusion-of-N (FusioN): a method that uses a general LLM judge to synthesize the most informative elements of each sample into a single final answer. We compare FusioN to BoN in two settings, (i) test-time scaling, where we sample and aggregate from a single model at test-time (ii) synthetic data generation, where we fuse samples from a pool of diverse teachers to improve a student model. We extensively benchmark both setups across 11 languages, 3 diverse tasks and varying model scales. Across the bench, FusioN consistently outperforms BoN showing versatility and robustness both in test-time scaling and in downstream gains from synthetic data generation. We also perform extensive analysis on FusioN, where it shows surprising strengths and robustness under challenging settings. These results show that we should shift how we think about evaluating and utilizing LLM generations from a monolithic measure of quality, to embracing their polylithic nature. This shift allows us to integrate diverse strengths, unlock latent potential, and achieve improvements that were previously inaccessible through selection alone.
Authors: Eileen Pan, Anna Seo Gyeong Choi, Maartje ter Hoeve, Skyler Seto, Allison Koenecke
Abstract: Large language models (LLMs) are ubiquitous in modern day natural language processing. However, previous work has shown degraded LLM performance for under-represented English dialects. We analyze the effects of typifying "standard" American English language questions as non-"standard" dialectal variants on multiple choice question answering tasks and find up to a 20% reduction in accuracy. Additionally, we investigate the grammatical basis of under-performance in non-"standard" English questions. We find that individual grammatical rules have varied effects on performance, but some are more consequential than others: three specific grammar rules (existential "it", zero copula, and y'all) can explain the majority of performance degradation observed in multiple dialects. We call for future work to investigate bias mitigation methods focused on individual, high-impact grammatical structures.
Authors: Ruqian Zhang, Yijiao Zhang, Juan Shen, Zhongyi Zhu, Annie Qu
Abstract: Large language models have made revolutionary progress in generating human-like text, yet their outputs often tend to be generic, exhibiting insufficient structural diversity, which limits personalized expression. Recent advances in diffusion models have opened new opportunities for improving language generation beyond the limitations of autoregressive paradigms. In this work, we propose a syntax-guided diffusion language model that integrates structural supervision and personalized conditioning to enhance text quality, diversity, and controllability. We introduce a cascaded framework that generates syntactic guidance before conditional text generation, and further generalize it to a novel noncascaded architecture for better alignment between structure and content. By incorporating syntactic information in the generating process, the proposed model better captures the lexical and structural characteristics of stylistic sentence construction. To enable fine-grained personalization, we develop a shared representation mechanism that facilitates information integration across users, supporting both faithful stylistic generation and generalizable zero-shot inference. Extensive experiments on multiple tasks demonstrate the superiority of our approach in fluency, diversity, and stylistic fidelity. Further qualitative analyses highlight its interpretability and flexibility in learning personalized patterns.
Authors: Nils Feldhus, Laura Kopf
Abstract: Understanding the decision-making processes of neural networks is a central goal of mechanistic interpretability. In the context of Large Language Models (LLMs), this involves uncovering the underlying mechanisms and identifying the roles of individual model components such as neurons and attention heads, as well as model abstractions such as the learned sparse features extracted by Sparse Autoencoders (SAEs). A rapidly growing line of work tackles this challenge by using powerful generator models to produce open-vocabulary, natural language concept descriptions for these components. In this paper, we provide the first survey of the emerging field of concept descriptions for model components and abstractions. We chart the key methods for generating these descriptions, the evolving landscape of automated and human metrics for evaluating them, and the datasets that underpin this research. Our synthesis reveals a growing demand for more rigorous, causal evaluation. By outlining the state of the art and identifying key challenges, this survey provides a roadmap for future research toward making models more transparent.
Authors: Samin Mahdipour Aghabagher, Saeedeh Momtazi
Abstract: Dialogue State Tracking (DST) is an essential element of conversational AI with the objective of deeply understanding the conversation context and leading it toward answering user requests. Due to high demands for open-domain and multi-turn chatbots, the traditional rule-based DST is not efficient enough, since it cannot provide the required adaptability and coherence for human-like experiences in complex conversations. This study proposes a hybrid DST model that utilizes rule-based methods along with language models, including BERT for slot filling and intent detection, XGBoost for intent validation, GPT for DST, and online agents for real-time answer generation. This model is uniquely designed to be evaluated on a comprehensive Persian multi-turn dialogue dataset and demonstrated significantly improved accuracy and coherence over existing methods in Persian-based chatbots. The results demonstrate how effectively a hybrid approach may improve DST capabilities, paving the way for conversational AI systems that are more customized, adaptable, and human-like.
Authors: Haonan Wang, Junfeng Sun, Mingjia Zhao, Wei Liu
Abstract: This article addresses embodied intelligence and reinforcement learning integration in the field of text processing, aiming to enhance text handling with more intelligence on the basis of embodied intelligence's perception and action superiority and reinforcement learning's decision optimization capability. Through detailed theoretical explanation and experimental exploration, a novel integration model is introduced. This model has been demonstrated to be very effective in a wide range oftext processing tasks, validating its applicative potential
Authors: Sukairaj Hafiz Imam, Tadesse Destaw Belay, Kedir Yassin Husse, Ibrahim Said Ahmad, Idris Abdulmumin, Hadiza Ali Umar, Muhammad Yahuza Bello, Joyce Nakatumba-Nabende, Seid Muhie Yimam, Shamsuddeen Hassan Muhammad
Abstract: ASR has achieved remarkable global progress, yet African low-resource languages remain rigorously underrepresented, producing barriers to digital inclusion across the continent with more than +2000 languages. This systematic literature review (SLR) explores research on ASR for African languages with a focus on datasets, models and training methods, evaluation techniques, challenges, and recommends future directions. We employ the PRISMA 2020 procedures and search DBLP, ACM Digital Library, Google Scholar, Semantic Scholar, and arXiv for studies published between January 2020 and July 2025. We include studies related to ASR datasets, models or metrics for African languages, while excluding non-African, duplicates, and low-quality studies (score <3/5). We screen 71 out of 2,062 records and we record a total of 74 datasets across 111 languages, encompassing approximately 11,206 hours of speech. Fewer than 15% of research provided reproducible materials, and dataset licensing is not clear. Self-supervised and transfer learning techniques are promising, but are hindered by limited pre-training data, inadequate coverage of dialects, and the availability of resources. Most of the researchers use Word Error Rate (WER), with very minimal use of linguistically informed scores such as Character Error Rate (CER) or Diacritic Error Rate (DER), and thus with limited application in tonal and morphologically rich languages. The existing evidence on ASR systems is inconsistent, hindered by issues like dataset availability, poor annotations, licensing uncertainties, and limited benchmarking. Nevertheless, the rise of community-driven initiatives and methodological advancements indicates a pathway for improvement. Sustainable development for this area will also include stakeholder partnership, creation of ethically well-balanced datasets, use of lightweight modelling techniques, and active benchmarking.
Authors: David Anugraha, Shou-Yi Hung, Zilu Tang, Annie En-Shiun Lee, Derry Tanti Wijaya, Genta Indra Winata
Abstract: Evaluation using Large Language Model (LLM) judges has been widely adopted in English and shown to be effective for automatic evaluation. However, their performance does not generalize well to non-English settings, and it remains unclear what constitutes effective multilingual training for such judges. In this paper, we introduce mR3, a massively multilingual, rubric-agnostic reward reasoning model trained on 72 languages, achieving the broadest language coverage in reward modeling to date. We present a comprehensive study of data and curriculum selection for training to identify effective strategies and data sources for building high-quality reward models, including the integration of target-language reasoning datasets. Our approach attains state-of-the-art performance on multilingual reward model benchmarks, surpassing much larger models (i.e., GPT-OSS-120B) while being up to 9x smaller, and its effectiveness is further confirmed through extensive ablation studies. Our models, data, and code are available as open source at https://github.com/rubricreward/mr3.
Authors: Mustafa Omer Gul, Claire Cardie, Tanya Goyal
Abstract: LLMs cannot reliably recognize their parametric knowledge boundaries and often hallucinate answers to outside-of-boundary questions. In contrast, humans recognize their limitations and can either seek external help for such questions or abstain. In this paper, we introduce MASH (Modeling Abstention via Selective Help-seeking), a training framework that readily extracts abstentions from LLMs. Our key idea is that any external help-seeking by an LLM, i.e. search tool use, can serve as a proxy for abstention if the external help (search) is appropriately penalized while simultaneously rewarding answer accuracy. MASH operationalizes this idea using reinforcement learning with a pay-per-search reward. We run experiments on three knowledge-intensive QA datasets. Our results show that MASH substantially improves upon the selective help-seeking performance of prior efficient search approaches; on multi-hop datasets, MASH improves answer accuracy by 7.6%. Furthermore, MASH demonstrates strong off-the-shelf abstention -- it can distinguish between unanswerable/answerable questions and selectively generate responses for answerable questions -- showcasing behavior analogous to specialized abstention approaches. We emphasize that contrary to prior abstention methods, MASH does not require pre-determining knowledge boundaries to construct training data. Instead, MASH's abstentions are a by-product of training for the auxiliary selective help-seeking task. Overall, we show that MASH training effectively aligns search tool use with parametric knowledge, which can be successfully leveraged for making abstention decisions.
Authors: Alexandrine Fortier, Thomas Thebaud, Jes\'us Villalba, Najim Dehak, Patrick Cardinal
Abstract: Large Language Models (LLMs) and their multimodal extensions are becoming increasingly popular. One common approach to enable multimodality is to cascade domain-specific encoders with an LLM, making the resulting model inherit vulnerabilities from all of its components. In this work, we present the first systematic study of audio backdoor attacks against speech language models. We demonstrate its effectiveness across four speech encoders and three datasets, covering four tasks: automatic speech recognition (ASR), speech emotion recognition, and gender and age prediction. The attack consistently achieves high success rates, ranging from 90.76% to 99.41%. To better understand how backdoors propagate, we conduct a component-wise analysis to identify the most vulnerable stages of the pipeline. Finally, we propose a fine-tuning-based defense that mitigates the threat of poisoned pretrained encoders.
Authors: Zhengliang Shi, Ruotian Ma, Jen-tse Huang, Xinbei Ma, Xingyu Chen, Mengru Wang, Qu Yang, Yue Wang, Fanghua Ye, Ziyang Chen, Shanyi Wang, Cixing Li, Wenxuan Wang, Zhaopeng Tu, Xiaolong Li, Zhaochun Ren, Linus
Abstract: Large language models (LLMs) are increasingly entrusted with high-stakes decisions that affect human welfare. However, the principles and values that guide these models when distributing scarce societal resources remain largely unexamined. To address this, we introduce the Social Welfare Function (SWF) Benchmark, a dynamic simulation environment where an LLM acts as a sovereign allocator, distributing tasks to a heterogeneous community of recipients. The benchmark is designed to create a persistent trade-off between maximizing collective efficiency (measured by Return on Investment) and ensuring distributive fairness (measured by the Gini coefficient). We evaluate 20 state-of-the-art LLMs and present the first leaderboard for social welfare allocation. Our findings reveal three key insights: (i) A model's general conversational ability, as measured by popular leaderboards, is a poor predictor of its allocation skill. (ii) Most LLMs exhibit a strong default utilitarian orientation, prioritizing group productivity at the expense of severe inequality. (iii) Allocation strategies are highly vulnerable, easily perturbed by output-length constraints and social-influence framing. These results highlight the risks of deploying current LLMs as societal decision-makers and underscore the need for specialized benchmarks and targeted alignment for AI governance.
Authors: Oussama Gabouj, Kamel Charaf, Ivan Zakazov, Nicolas Baldwin, Robert West
Abstract: Large Language Models (LLMs) achieve strong performance across diverse tasks, but their effectiveness often depends on the quality of the provided context. Retrieval-Augmented Generation (RAG) enriches prompts with external information, but its reliance on static databases constrains adaptability and can result in irrelevant demonstrations. In this work, we propose a Generative Retrieval-Aligned Demonstrator (GRAD), a dynamic demonstration-based approach where an LLM model is trained to generate input-specific concise demonstrations. By tailoring demonstrations to each input, our method offers better contextual support than traditional RAG approaches. We demonstrate the superiority of GRAD under budget constraints, where we limit both the number of tokens used per demonstration and the number of tokens used for the final output. Trained solely on a math dataset, GRAD consistently outperforms strong baselines on Qwen2.5-14B across mathematical reasoning and advanced STEM questions, highlighting GRAD's robust generalization to out-of-distribution (OOD) domains such as physics, chemistry, and computer science. Furthermore, we show that demonstrations generated by trained smaller models can effectively guide larger target models, reducing training costs while maintaining competitive accuracy. Overall, this work introduces a scalable demonstration generator model presenting the first step toward a dynamic few-shot learning paradigm in resource-constrained settings. We release the code used for the project.
Authors: Jiayi Zhang, Simon Yu, Derek Chong, Anthony Sicilia, Michael R. Tomz, Christopher D. Manning, Weiyan Shi
Abstract: Post-training alignment often reduces LLM diversity, leading to a phenomenon known as mode collapse. Unlike prior work that attributes this effect to algorithmic limitations, we identify a fundamental, pervasive data-level driver: typicality bias in preference data, whereby annotators systematically favor familiar text as a result of well-established findings in cognitive psychology. We formalize this bias theoretically, verify it on preference datasets empirically, and show that it plays a central role in mode collapse. Motivated by this analysis, we introduce Verbalized Sampling, a simple, training-free prompting strategy to circumvent mode collapse. VS prompts the model to verbalize a probability distribution over a set of responses (e.g., ``Generate 5 jokes about coffee and their corresponding probabilities''). Comprehensive experiments show that VS significantly improves performance across creative writing (poems, stories, jokes), dialogue simulation, open-ended QA, and synthetic data generation, without sacrificing factual accuracy and safety. For instance, in creative writing, VS increases diversity by 1.6-2.1x over direct prompting. We further observe an emergent trend that more capable models benefit more from VS. In sum, our work provides a new data-centric perspective on mode collapse and a practical inference-time remedy that helps unlock pre-trained generative diversity.
Authors: Qingyuan Liu, Jia-Chen Gu, Yunzhi Yao, Hong Wang, Nanyun Peng
Abstract: Large language models (LLMs) require constant updates to remain aligned with evolving real-world knowledge. Model editing offers a lightweight alternative to retraining, but sequential editing often destabilizes representations and induces catastrophic forgetting. In this work, we seek to better understand and mitigate performance degradation caused by sequential editing. We hypothesize that hyperspherical uniformity, a property that maintains uniform distribution of neuron weights on a hypersphere, helps the model remain stable, retain prior knowledge, while still accommodate new updates. We use Hyperspherical Energy (HE) to quantify neuron uniformity during editing, and examine its correlation with editing performance. Empirical studies across widely used editing methods reveals a strong correlation between HE dynamics and editing performance, with editing failures consistently coinciding with high HE fluctuations. We further theoretically prove that HE dynamics impose a lower bound on the degradation of pretrained knowledge, highlighting why HE stability is crucial for knowledge retention. Motivated by these insights, we propose SPHERE (Sparse Projection for Hyperspherical Energy-Regularized Editing), an HE-driven regularization strategy that stabilizes neuron weight distributions, ultimately preserving prior knowledge while enabling reliable sequential updates. Specifically, SPHERE identifies a sparse space complementary to the principal hyperspherical directions of the pretrained weight matrices and projects new knowledge onto it, attenuating perturbations on the principal directions. Extensive experiments on LLaMA3 (8B) and Qwen2.5 (7B) show that SPHERE outperforms the best baseline in editing capability by an average of 16.41%, while most faithfully preserving general model performance, thereby offering a principled path toward reliable large-scale knowledge editing.
Authors: Kajwan Ziaoddini
Abstract: This paper examines how musical symbolism is produced and circulated in online communities by combining content-based music analysis with a lightweight network perspective on lyrics. Using a curated corpus of 275 chart-topping songs enriched with audio descriptors (energy, danceability, loudness, liveness, valence, acousticness, speechiness, popularity) and full lyric transcripts, we build a reproducible pipeline that (i) quantifies temporal trends in sonic attributes, (ii) models lexical salience and co-occurrence, and (iii) profiles mood by genre. We find a decade-long decline in energy (79 -> 58) alongside a rise in danceability (59 -> 73); valence peaks in 2013 (63) and dips in 2014-2016 (42) before partially recovering. Correlation analysis shows strong coupling of energy with loudness (r = 0.74) and negative associations for acousticness with both energy (r = -0.54) and loudness (r = -0.51); danceability is largely orthogonal to other features (|r| < 0.20). Lyric tokenization (>114k tokens) reveals a pronoun-centric lexicon "I/you/me/my" and a dense co-occurrence structure in which interpersonal address anchors mainstream narratives. Mood differs systematically by style: R&B exhibits the highest mean valence (96), followed by K-Pop/Pop (77) and Indie/Pop (70), whereas Latin/Reggaeton is lower (37) despite high danceability. Read through a subcultural identity lens, these patterns suggest the mainstreaming of previously peripheral codes and a commercial preference for relaxed yet rhythmically engaging productions that sustain collective participation without maximal intensity. Methodologically, we contribute an integrated MIR-plus-network workflow spanning summary statistics, correlation structure, lexical co-occurrence matrices, and genre-wise mood profiling that is robust to modality sparsity and suitable for socially aware recommendation or community-level diffusion studies.
Authors: Alvaro Vallejo Ram\'irez
Abstract: Purpose. This study analyzes the digital representation of the Iran-Israel conflict that occurred in June 2025, based on 120,000 comments posted on YouTube. It sought to identify discursive positions regarding the actors involved and to examine how media and algorithmic biases shape digital conversations. Methodology. A mixed-methods design with triangulation was adopted. In the quantitative phase, natural language processing techniques and machine learning models (BERT and XLM-RoBERTa) were used to classify comments into ten categories. In the qualitative phase, a critical analysis of media context and ideological narratives was conducted, complemented by manual annotation and supervised training. This strategy enabled the integration of statistical robustness with contextual understanding. Results and conclusions. The findings reveal a clear overrepresentation of pro-Palestinian and anti-United States/Israel discourses, while pro-United States and anti-Palestinian positions were marginal. Iran, usually rendered invisible in global media, emerged as a central actor in the digital conversation during the conflict, suggesting a narrative shift away from previous hegemonic frameworks. Likewise, the results confirm the influence of algorithmic biases in amplifying certain discourses while limiting others. Original contributions. This work combines computational analysis and philosophical critique for the study of digital controversies, providing a methodological framework replicable in geopolitical contexts. It is one of the first Spanish-language studies to map, through artificial intelligence and critical analysis, discourses on an international conflict on YouTube, highlighting asymmetries and narrative disputes that are often overlooked.
Authors: Ziyi Zeng, Zhenyang Cai, Yixi Cai, Xidong Wang, Junying Chen, Rongsheng Wang, Yipeng Liu, Siqi Cai, Benyou Wang, Zhiguo Zhang, Haizhou Li
Abstract: Electroencephalography (EEG) interpretation using multimodal large language models (MLLMs) offers a novel approach for analyzing brain signals. However, the complex nature of brain activity introduces critical challenges: EEG signals simultaneously encode both cognitive processes and intrinsic neural states, creating a mismatch in EEG paired-data modality that hinders effective cross-modal representation learning. Through a pivot investigation, we uncover complementary relationships between these modalities. Leveraging this insight, we propose mapping EEG signals and their corresponding modalities into a unified semantic space to achieve generalized interpretation. To fully enable conversational capabilities, we further introduce WaveMind-Instruct-338k, the first cross-task EEG dataset for instruction tuning. The resulting model demonstrates robust classification accuracy while supporting flexible, open-ended conversations across four downstream tasks, thereby offering valuable insights for both neuroscience research and the development of general-purpose EEG models.
Authors: Gregory D. Baker, Scott McCallum, Dirk Pattinson
Abstract: Many real-world machine learning problems involve inherently hierarchical data, yet traditional approaches rely on Euclidean metrics that fail to capture the discrete, branching nature of hierarchical relationships. We present a theoretical foundation for machine learning in p-adic metric spaces, which naturally respect hierarchical structure. Our main result proves that an n-dimensional plane minimizing the p-adic sum of distances to points in a dataset must pass through at least n + 1 of those points -- a striking contrast to Euclidean regression that highlights how p-adic metrics better align with the discrete nature of hierarchical data. As a corollary, a polynomial of degree n constructed to minimise the p-adic sum of residuals will pass through at least n + 1 points. As a further corollary, a polynomial of degree n approximating a higher degree polynomial at a finite number of points will yield a difference polynomial that has distinct rational roots. We demonstrate the practical significance of this result through two applications in natural language processing: analyzing hierarchical taxonomies and modeling grammatical morphology. These results suggest that p-adic metrics may be fundamental to properly handling hierarchical data structures in machine learning. In hierarchical data, interpolation between points often makes less sense than selecting actual observed points as representatives.
Authors: Dongqi Zheng
Abstract: Large Reasoning Language Models (LRLMs or LRMs) demonstrate remarkable capabilities in complex reasoning tasks, but suffer from significant computational inefficiencies due to overthinking phenomena. Existing efficient reasoning methods face the challenge of balancing reasoning quality with inference cost reduction. We propose \textbf{Adaptive Reasoning Suppression (ARS)}, a novel training-free approach that dynamically suppresses redundant reasoning steps while preserving accuracy through adaptive certainty monitoring. ARS introduces a multi-checkpoint certainty estimation mechanism with progressive suppression thresholds, achieving superior efficiency compared to static suppression methods. Our extensive evaluation across mathematical reasoning benchmarks using multiple model architectures demonstrates that ARS achieves up to 53%, 46.1%, and 57.9% in token, latency and energy reduction, while maintaining or improving accuracy.
Authors: Nima Sheikholeslami, Erfan Hosseini, Patrice Bechard, Srivatsava Daruru, Sai Rajeswar
Abstract: Dual-encoder retrievers depend on the principle that relevant documents should score higher than irrelevant ones for a given query. Yet the dominant Noise Contrastive Estimation (NCE) objective, which underpins Contrastive Loss, optimizes a softened ranking surrogate that we rigorously prove is fundamentally oblivious to score separation quality and unrelated to AUC. This mismatch leads to poor calibration and suboptimal performance in downstream tasks like retrieval-augmented generation (RAG). To address this fundamental limitation, we introduce the MW loss, a new training objective that maximizes the Mann-Whitney U statistic, which is mathematically equivalent to the Area under the ROC Curve (AUC). MW loss encourages each positive-negative pair to be correctly ranked by minimizing binary cross entropy over score differences. We provide theoretical guarantees that MW loss directly upper-bounds the AoC, better aligning optimization with retrieval goals. We further promote ROC curves and AUC as natural threshold free diagnostics for evaluating retriever calibration and ranking quality. Empirically, retrievers trained with MW loss consistently outperform contrastive counterparts in AUC and standard retrieval metrics. Our experiments show that MW loss is an empirically superior alternative to Contrastive Loss, yielding better-calibrated and more discriminative retrievers for high-stakes applications like RAG.
Authors: Houjun Liu, Shikhar Murty, Christopher D. Manning, R\'obert Csord\'as
Abstract: Current approaches for scaling inference-time compute in transformers rely on training them to emit explicit chain-of-thought tokens before producing an answer. While these methods are powerful, they are limited because they cannot be applied during pretraining and are limited to only serially-generated, natural-language verbalization to scale inference-time compute. In this work, we propose Thoughtbubbles, a transformer variant that natively performs parallel adaptive computation in latent space by learning to fork or delete residual streams. Thus, tokens that require a large amount of computation can form a "bubble" of cloned residuals in the middle of the network for additional thinking. Crucially, this behavior is learned during pretraining with only language modeling loss. Thoughtbubbles outperforms both standard decoder LMs as well as non-adaptive parallel computation approaches on OpenWebText and peS2o perplexity and in zero-shot evaluations such as HellaSwag and LAMBADA after pretraining across 150M to 772M parameter scales. The implicit nature of our method enables adaptive computation to be learned starting at pretraining time, paving the way to unify train and test-time behavior for reasoning models.
Authors: Priyank Dubey
Abstract: Link prediction is one of the fundamental problems in graph theory, critical for understanding and forecasting the evolution of complex systems like social and biological networks. While classical heuristics capture certain aspects of graph topology, they often struggle to optimally integrate local and global structural information or adapt to complex dependencies. Quantum computing offers a powerful alternative by leveraging superposition for simultaneous multi-path exploration and interference-driven integration of both local and global graph features. In this work, we introduce QSearchNet, a quantum-inspired framework based on Discrete-Time Quantum Walk (DTQW) dynamics and Grover's amplitude amplification. QSearchNet simulates a topology-aware quantum evolution to propagate amplitudes across multiple nodes simultaneously. By aligning interference patterns through quantum reflection and oracle-like phase-flip operation, it adaptively prioritizes multi-hop dependencies and amplifies structurally relevant paths corresponding to potential connections. Experiments on diverse real-world networks demonstrate competitive performance, particularly with hard negative samples under realistic evaluation conditions.
Authors: T. James Brandt
Abstract: Adaptive chatbots that mimic a user's linguistic style can build rapport and engagement, yet unconstrained mimicry risks an agent that feels unstable or sycophantic. We present a computational evaluation framework that makes the core design tension explicit: balancing moment-to-moment linguistic synchrony against long-term persona stability. Using an 8-dimensional style vector and a closed-loop "base+delta" prompting architecture, we simulate and compare explicit adaptation policies - Uncapped, Cap, Exponential Moving Average (EMA), Dead-Band, and Hybrids - on a human-log dataset. Our analysis maps a clear Pareto frontier: bounded policies achieve substantial gains in stability at a modest cost to synchrony. For example, a Hybrid (EMA+Cap) raises stability from 0.542 to 0.878 (+62%) while reducing synchrony by only 17%. We confirm this trade-off through large-scale replications on three public corpora (DailyDialog, Persona-Chat, EmpatheticDialogues) and LLM-in-the-loop validation across two model families. Furthermore, we quantify "prompt legibility," showing that frontier policies reduce instruction churn and cut jarring register flips (major tone changes) from 0.254 to 0.092, yielding systems that are easier to reason about and maintain. Taken together, our framework provides a general evaluation harness for style adaptation; a systematic ablation that identifies Pareto-efficient policies; robust validation across diverse datasets and models; and novel legibility metrics linking policy choices to system maintainability.
Authors: Minseok Jeon, Seunghyun Park
Abstract: We present GDLNN, a new graph machine learning architecture, for graph classification tasks. GDLNN combines a domain-specific programming language, called GDL, with neural networks. The main strength of GDLNN lies in its GDL layer, which generates expressive and interpretable graph representations. Since the graph representation is interpretable, existing model explanation techniques can be directly applied to explain GDLNN's predictions. Our evaluation shows that the GDL-based representation achieves high accuracy on most graph classification benchmark datasets, outperforming dominant graph learning methods such as GNNs. Applying an existing model explanation technique also yields high-quality explanations of GDLNN's predictions. Furthermore, the cost of GDLNN is low when the explanation cost is included.
Authors: Xudong Zhu, Mohammad Mahdi Khalili, Zhihui Zhu
Abstract: Sparse autoencoders (SAEs) have emerged as powerful techniques for interpretability of large language models (LLMs), aiming to decompose hidden states into meaningful semantic features. While several SAE variants have been proposed, there remains no principled framework to derive SAEs from the original dictionary learning formulation. In this work, we introduce such a framework by unrolling the proximal gradient method for sparse coding. We show that a single-step update naturally recovers common SAE variants, including ReLU, JumpReLU, and TopK. Through this lens, we reveal a fundamental limitation of existing SAEs: their sparsity-inducing regularizers enforce non-negativity, preventing a single feature from representing bidirectional concepts (e.g., male vs. female). This structural constraint fragments semantic axes into separate, redundant features, limiting representational completeness. To address this issue, we propose AbsTopK SAE, a new variant derived from the $\ell_0$ sparsity constraint that applies hard thresholding over the largest-magnitude activations. By preserving both positive and negative activations, AbsTopK uncovers richer, bidirectional conceptual representations. Comprehensive experiments across four LLMs and seven probing and steering tasks show that AbsTopK improves reconstruction fidelity, enhances interpretability, and enables single features to encode contrasting concepts. Remarkably, AbsTopK matches or even surpasses the Difference-in-Mean method, a supervised approach that requires labeled data for each concept and has been shown in prior work to outperform SAEs.
Authors: Sarvesh Soni, Dina Demner-Fushman
Abstract: Automated approaches to answer patient-posed health questions are rising, but selecting among systems requires reliable evaluation. The current gold standard for evaluating the free-text artificial intelligence (AI) responses--human expert review--is labor-intensive and slow, limiting scalability. Automated metrics are promising yet variably aligned with human judgments and often context-dependent. To address the feasibility of automating the evaluation of AI responses to hospitalization-related questions posed by patients, we conducted a large systematic study of evaluation approaches. Across 100 patient cases, we collected responses from 28 AI systems (2800 total) and assessed them along three dimensions: whether a system response (1) answers the question, (2) appropriately uses clinical note evidence, and (3) uses general medical knowledge. Using clinician-authored reference answers to anchor metrics, automated rankings closely matched expert ratings. Our findings suggest that carefully designed automated evaluation can scale comparative assessment of AI systems and support patient-clinician communication.
Authors: Yen-Shan Chen, Sian-Yao Huang, Cheng-Lin Yang, Yun-Nung Chen
Abstract: Existing data poisoning attacks on retrieval-augmented generation (RAG) systems scale poorly because they require costly optimization of poisoned documents for each target phrase. We introduce Eyes-on-Me, a modular attack that decomposes an adversarial document into reusable Attention Attractors and Focus Regions. Attractors are optimized to direct attention to the Focus Region. Attackers can then insert semantic baits for the retriever or malicious instructions for the generator, adapting to new targets at near zero cost. This is achieved by steering a small subset of attention heads that we empirically identify as strongly correlated with attack success. Across 18 end-to-end RAG settings (3 datasets $\times$ 2 retrievers $\times$ 3 generators), Eyes-on-Me raises average attack success rates from 21.9 to 57.8 (+35.9 points, 2.6$\times$ over prior work). A single optimized attractor transfers to unseen black box retrievers and generators without retraining. Our findings establish a scalable paradigm for RAG data poisoning and show that modular, reusable components pose a practical threat to modern AI systems. They also reveal a strong link between attention concentration and model outputs, informing interpretability research.
Authors: Minki Kang, Wei-Ning Chen, Dongge Han, Huseyin A. Inan, Lukas Wutschitz, Yanzhi Chen, Robert Sim, Saravan Rajmohan
Abstract: Large language models (LLMs) are increasingly deployed as agents in dynamic, real-world environments, where success requires both reasoning and effective tool use. A central challenge for agentic tasks is the growing context length, as agents must accumulate long histories of actions and observations. This expansion raises costs and reduces efficiency in long-horizon tasks, yet prior work on context compression has mostly focused on single-step tasks or narrow applications. We introduce Agent Context Optimization (ACON), a unified framework that optimally compresses both environment observations and interaction histories into concise yet informative condensations. ACON leverages compression guideline optimization in natural language space: given paired trajectories where full context succeeds but compressed context fails, capable LLMs analyze the causes of failure, and the compression guideline is updated accordingly. Furthermore, we propose distilling the optimized LLM compressor into smaller models to reduce the overhead of the additional module. Experiments on AppWorld, OfficeBench, and Multi-objective QA show that ACON reduces memory usage by 26-54% (peak tokens) while largely preserving task performance, preserves over 95% of accuracy when distilled into smaller compressors, and enhances smaller LMs as long-horizon agents with up to 46% performance improvement.
Authors: Rosni Vasu, Peter Jansen, Pao Siangliulue, Cristina Sarasua, Abraham Bernstein, Peter Clark, Bhavana Dalvi Mishra
Abstract: While there has been a surge of interest in automated scientific discovery (ASD), especially with the emergence of LLMs, it remains challenging for tools to generate hypotheses that are both testable and grounded in the scientific literature. Additionally, existing ideation tools are not adaptive to prior experimental outcomes. We developed HARPA to address these challenges by incorporating the ideation workflow inspired by human researchers. HARPA first identifies emerging research trends through literature mining, then explores hypothesis design spaces, and finally converges on precise, testable hypotheses by pinpointing research gaps and justifying design choices. Our evaluations show that HARPA-generated hypothesis-driven research proposals perform comparably to a strong baseline AI-researcher across most qualitative dimensions (e.g., specificity, novelty, overall quality), but achieve significant gains in feasibility(+0.78, p$<0.05$, bootstrap) and groundedness (+0.85, p$<0.01$, bootstrap) on a 10-point Likert scale. When tested with the ASD agent (CodeScientist), HARPA produced more successful executions (20 vs. 11 out of 40) and fewer failures (16 vs. 21 out of 40), showing that expert feasibility judgments track with actual execution success. Furthermore, to simulate how researchers continuously refine their understanding of what hypotheses are both testable and potentially interesting from experience, HARPA learns a reward model that scores new hypotheses based on prior experimental outcomes, achieving approx. a 28\% absolute gain over HARPA's untrained baseline scorer. Together, these methods represent a step forward in the field of AI-driven scientific discovery.
Authors: Chen-An Li, Tzu-Han Lin, Hung-yi Lee
Abstract: Large audio-language models (LALMs) unify speech and text processing, but their robustness in noisy real-world settings remains underexplored. We investigate how irrelevant audio, such as silence, synthetic noise, and environmental sounds, affects text reasoning tasks where audio is unnecessary. Across three text-based benchmarks, we find that even non-informative audio reduces accuracy and increases prediction volatility; the severity of interference scales with longer durations, higher amplitudes, and elevated decoding temperatures. Silence, often assumed neutral, destabilizes outputs as strongly as synthetic noise. While larger models show greater resilience, vulnerabilities persist across all evaluated systems. We further test mitigation strategies and find that prompting shows limited effectiveness, whereas self-consistency improves stability at the cost of increased computation. Our results reveal cross-modal interference as a key robustness challenge and highlight the need for efficient fusion strategies that preserve reasoning performance in the presence of irrelevant inputs.
Authors: Yu-Xiang Lin, Chen-An Li, Sheng-Lun Wei, Po-Chun Chen, Hsin-Hsi Chen, Hung-yi Lee
Abstract: Large audio-language models (LALMs) are often used in tasks that involve reasoning over ordered options. An open question is whether their predictions are influenced by the order of answer choices, which would indicate a form of selection bias and undermine their reliability. In this paper, we identify and analyze this problem in LALMs. We demonstrate that no model is immune to this bias through extensive experiments on six LALMs across three widely used benchmarks and their spoken counterparts. Shuffling the order of answer options can cause performance fluctuations of up to 24% and even change model rankings, raising concerns about the reliability of current evaluation practices. We also study permutation-based strategies and show that they can mitigate bias in most cases. Our work represents the first systematic investigation of this issue in LALMs, and we hope it raises awareness and motivates further research in this direction.
Authors: Alessio Devoto, Maximilian Jeblick, Simon J\'egou
Abstract: Memory consumption of the Key-Value (KV) cache represents a major bottleneck for efficient large language model inference. While attention-score-based KV cache pruning shows promise, it faces critical practical limitations: attention scores from future tokens are unavailable during compression, and modern implementations like Flash Attention do not materialize the full attention matrix, making past scores inaccessible. To overcome these challenges, we introduce $\textbf{Expected Attention, a training-free compression method}$ that estimates KV pairs importance by predicting how future queries will attend to them. Our approach leverages the distributional properties of LLM activations to compute expected attention scores in closed form for each KV pair. These scores enable principled ranking and pruning of KV pairs with minimal impact on the residual stream, achieving effective compression without performance degradation. Importantly, our method operates seamlessly across both prefilling and decoding phases, consistently outperforming state-of-the-art baselines in both scenarios. Finally, $\textbf{we release KVPress, a comprehensive library to enable researchers to implement and benchmark KV cache compression methods, already including more than 20 techniques}$.
Authors: Thong Nguyen, Yibin Lei, Jia-Huei Ju, Eugene Yang, Andrew Yates
Abstract: Learned Sparse Retrieval (LSR) combines the efficiency of bi-encoders with the transparency of lexical matching, but existing approaches struggle to scale beyond English. We introduce MILCO, an LSR architecture that maps queries and documents from different languages into a shared English lexical space via a multilingual connector. MILCO is trained with a specialized two-stage regime that combines Sparse Alignment Pretraining with contrastive training to provide representation transparency and effectiveness while mitigating semantic collapse. Motivated by the observation that uncommon entities are often lost when projected into English, we propose a new LexEcho head, which enhances robustness by augmenting the English lexical representation with a source-language view obtained through a special [ECHO] token. MILCO achieves state-of-the-art multilingual and cross-lingual LSR performance, outperforming leading dense, sparse, and multi-vector baselines such as BGE-M3 and Qwen3-Embed on standard multilingual benchmarks, while supporting dynamic efficiency through post-hoc pruning. Notably, when using mass-based pruning to reduce document representations to only 30 active dimensions on average, MILCO 560M outperforms the similarly-sized Qwen3-Embed 0.6B with 1024 dimensions.
Authors: Nurbek Tastan, Samuel Horvath, Karthik Nandakumar
Abstract: Multi-agent systems (MAS) based on Large Language Models (LLMs) have the potential to solve tasks that are beyond the reach of any single LLM. However, this potential can only be realized when the collaboration mechanism between agents is optimized. Specifically, optimizing the communication structure between agents is critical for fruitful collaboration. Most existing approaches rely on fixed topologies, pretrained graph generators, optimization over edges, or employ external LLM judges, thereby adding to the complexity. In this work, we introduce a response-conditioned framework that adapts communication on-the-fly. Agents independently generate responses to the user query and assess peer contributions using an approximation of the Shapley value. A directed acyclic graph (DAG) is then constructed to regulate the propagation of the responses among agents, which ensures stable and efficient message transmission from high-contributing agents to others. This graph is dynamically updated based on the agent responses from the previous collaboration round. Since the proposed framework enables the self-organization of agents without additional supervision or training, we refer to it as SelfOrg. The SelfOrg framework goes beyond task- and query-level optimization and takes into account the stochastic nature of agent responses. Experiments with both strong and weak LLM backends demonstrate robust performance, with significant gains in the weak regime where prior methods collapse. We also theoretically show that multiple agents increase the chance of correctness and that the correct responses naturally dominate the information flow.
Authors: Yifei Cao, Changhao Jiang, Jiabao Zhuang, Jiajun Sun, Ming Zhang, Zhiheng Xi, Hui Li, Shihan Dou, Yuran Wang, Yunke Zhang, Tao Ji, Tao Gui, Qi Zhang, Xuanjing Huang
Abstract: Assessing the perceptual quality of synthetic speech is crucial for guiding the development and refinement of speech generation models. However, it has traditionally relied on human subjective ratings such as the Mean Opinion Score (MOS), which depend on manual annotations and often suffer from inconsistent rating standards and poor reproducibility. To address these limitations, we introduce MOS-RMBench, a unified benchmark that reformulates diverse MOS datasets into a preference-comparison setting, enabling rigorous evaluation across different datasets. Building on MOS-RMBench, we systematically construct and evaluate three paradigms for reward modeling: scalar reward models, semi-scalar reward models, and generative reward models (GRMs). Our experiments reveal three key findings: (1) scalar models achieve the strongest overall performance, consistently exceeding 74% accuracy; (2) most models perform considerably worse on synthetic speech than on human speech; and (3) all models struggle on pairs with very small MOS differences. To improve performance on these challenging pairs, we propose a MOS-aware GRM that incorporates an MOS-difference-based reward function, enabling the model to adaptively scale rewards according to the difficulty of each sample pair. Experimental results show that the MOS-aware GRM significantly improves fine-grained quality discrimination and narrows the gap with scalar models on the most challenging cases. We hope this work will establish both a benchmark and a methodological framework to foster more rigorous and scalable research in automatic speech quality assessment.
Authors: Divy Kala, Eshika Khandelwal, Makarand Tapaswi
Abstract: Audio descriptions (ADs) narrate important visual details in movies, enabling Blind and Low Vision (BLV) users to understand narratives and appreciate visual details. Existing works in automatic AD generation mostly focus on few-second trimmed clips, and evaluate them by comparing against a single ground-truth reference AD. However, writing ADs is inherently subjective. Through alignment and analysis of two independent AD tracks for the same movies, we quantify the subjectivity in when and whether to describe, and what and how to highlight. Thus, we show that working with trimmed clips is inadequate. We propose ADQA, a QA benchmark that evaluates ADs at the level of few-minute long, coherent video segments, testing whether they would help BLV users understand the story and appreciate visual details. ADQA features visual appreciation (VA) questions about visual facts and narrative understanding (NU) questions based on the plot. Through ADQA, we show that current AD generation methods lag far behind human-authored ADs. We conclude with several recommendations for future work and introduce a public leaderboard for benchmarking.
Authors: Maxime M\'eloux, Maxime Peyrard, Fran\c{c}ois Portet
Abstract: The development of trustworthy artificial intelligence requires moving beyond black-box performance metrics toward an understanding of models' internal computations. Mechanistic Interpretability (MI) aims to meet this need by identifying the algorithmic mechanisms underlying model behaviors. Yet, the scientific rigor of MI critically depends on the reliability of its findings. In this work, we argue that interpretability methods, such as circuit discovery, should be viewed as statistical estimators, subject to questions of variance and robustness. To illustrate this statistical framing, we present a systematic stability analysis of a state-of-the-art circuit discovery method: EAP-IG. We evaluate its variance and robustness through a comprehensive suite of controlled perturbations, including input resampling, prompt paraphrasing, hyperparameter variation, and injected noise within the causal analysis itself. Across a diverse set of models and tasks, our results demonstrate that EAP-IG exhibits high structural variance and sensitivity to hyperparameters, questioning the stability of its findings. Based on these results, we offer a set of best-practice recommendations for the field, advocating for the routine reporting of stability metrics to promote a more rigorous and statistically grounded science of interpretability.
Authors: Kevin Zhang, Kuangzhi Ge, Xiaowei Chi, Renrui Zhang, Shaojun Shi, Zhen Dong, Sirui Han, Shanghang Zhang
Abstract: Trained on internet-scale video data, generative world models are increasingly recognized as powerful world simulators that can generate consistent and plausible dynamics over structure, motion, and physics. This raises a natural question: with the advent of strong video foundational models, might they supplant conventional vision encoder paradigms for general-purpose multimodal understanding? While recent studies have begun to explore the potential of world models on common vision tasks, these explorations typically lack a systematic investigation of generic, multimodal tasks. In this work, we strive to investigate the capabilities when world model priors are transferred into Vision-Language Models: we re-purpose a video diffusion model as a generative encoder to perform a single denoising step and treat the resulting latents as a set of visual embedding. We empirically investigate this class of models, which we refer to as World-Language Models (WorldLMs), and we find that generative encoders can capture latents useful for downstream understanding that show distinctions from conventional encoders. Naming our best-performing variant Dynamic Vision Aligner (DyVA), we further discover that this method significantly enhances spatial reasoning abilities and enables single-image models to perform multi-frame reasoning. Through the curation of a suite of visual reasoning tasks, we find DyVA to surpass both open-source and proprietary baselines, achieving state-of-the-art or comparable performance. We attribute these gains to WorldLM's inherited motion-consistency internalization from video pre-training. Finally, we systematically explore extensive model designs to highlight promising directions for future work. We hope our study can pave the way for a new family of VLMs that leverage priors from world models and are on a promising path towards generalist vision learners.
Authors: Thiziri Nait Saada, Louis Bethune, Michal Klein, David Grangier, Marco Cuturi, Pierre Ablin
Abstract: Large-scale models are pretrained on massive web-crawled datasets containing documents of mixed quality, making data filtering essential. A popular method is Classifier-based Quality Filtering (CQF), which trains a binary classifier to distinguish between pretraining data and a small, high-quality set. It assigns each pretraining document a quality score defined as the classifier's score and retains only the top-scoring ones. We provide an in-depth analysis of CQF. We show that while CQF improves downstream task performance, it does not necessarily enhance language modeling on the high-quality dataset. We explain this paradox by the fact that CQF implicitly filters the high-quality dataset as well. We further compare the behavior of models trained with CQF to those trained on synthetic data of increasing quality, obtained via random token permutations, and find starkly different trends. Our results challenge the view that CQF captures a meaningful notion of data quality.
Authors: Roksana Goworek, Olivia Macmillan-Scott, Eda B. \"Ozyi\u{g}it
Abstract: Cross-lingual information retrieval (CLIR) addresses the challenge of retrieving relevant documents written in languages different from that of the original query. Research in this area has typically framed the task as monolingual retrieval augmented by translation, treating retrieval methods and cross-lingual capabilities in isolation. Both monolingual and cross-lingual retrieval usually follow a pipeline of query expansion, ranking, re-ranking and, increasingly, question answering. Recent advances, however, have shifted from translation-based methods toward embedding-based approaches and leverage multilingual large language models (LLMs), for which aligning representations across languages remains a central challenge. The emergence of cross-lingual embeddings and multilingual LLMs has introduced a new paradigm, offering improved retrieval performance and enabling answer generation. This survey provides a comprehensive overview of developments from early translation-based methods to state-of-the-art embedding-driven and generative techniques. It presents a structured account of core CLIR components, evaluation practices, and available resources. Persistent challenges such as data imbalance and linguistic variation are identified, while promising directions are suggested for advancing equitable and effective cross-lingual information retrieval. By situating CLIR within the broader landscape of information retrieval and multilingual language processing, this work not only reviews current capabilities but also outlines future directions for building retrieval systems that are robust, inclusive, and adaptable.
Authors: Yihong Wu, Liheng Ma, Lei Ding, Muzhi Li, Xinyu Wang, Kejia Chen, Zhan Su, Zhanguang Zhang, Chenyang Huang, Yingxue Zhang, Mark Coates, Jian-Yun Nie
Abstract: Group Relative Policy Optimization (GRPO) is a prominent reinforcement learning algorithm for post-training Large Language Models (LLMs). It is commonly believed that GRPO necessitates a large group size to ensure stable training via precise statistical estimation, which incurs substantial computational overhead. In this work, we challenge this assumption by reframing GRPO as a form of contrastive learning, which reveals a fundamental connection to Direct Preference Optimization (DPO). Motivated by DPO's empirical success, we investigate the minimal two-rollout case (2-GRPO), a configuration previously deemed infeasible. We provide a rigorous theoretical analysis to validate 2-GRPO and demonstrate empirically that it achieves performance on par with 16-GRPO, despite using only 1/8 of the rollouts and reducing training time by over 70%.
Authors: Emiru Tsunoo, Hayato Futami, Yosuke Kashiwagi, Siddhant Arora, Shinji Watanabe
Abstract: For streaming speech recognition, a Transformer-based encoder has been widely used with block processing. Although many studies addressed improving emission latency of transducers, little work has been explored for improving encoding latency of the block processing. We seek to reduce latency by frequently emitting a chunk with a small shift rather than scarce large-chunk emissions, resulting in higher computational costs. To efficiently compute with the small chunk shift, we propose a new encoder, Spiralformer, tailored for block processing by combining layer dropping and early exiting. We skip layer computation in a cyclic manner and shift the computed layer in each block spirally, which completes computation for all the layers over the block processing. Experimentally, we observed that our method achieved 21.6% reduction in the averaged token emission delay in Librispeech, and 7.0% in CSJ, compared with the baseline with similar computational cost and word error rates.
Authors: Boshi Wang, Weijian Xu, Yunsheng Li, Mei Gao, Yujia Xie, Huan Sun, Dongdong Chen
Abstract: Code localization is a fundamental challenge in repository-level software engineering tasks such as bug fixing. While existing methods equip language agents with comprehensive tools/interfaces to fetch information from the repository, they overlook the critical aspect of memory, where each instance is typically handled from scratch assuming no prior repository knowledge. In contrast, human developers naturally build long-term repository memory, such as the functionality of key modules and associations between various bug types and their likely fix locations. In this work, we augment language agents with such memory by leveraging a repository's commit history - a rich yet underutilized resource that chronicles the codebase's evolution. We introduce tools that allow the agent to retrieve from a non-parametric memory encompassing recent historical commits and linked issues, as well as functionality summaries of actively evolving parts of the codebase identified via commit patterns. We demonstrate that augmenting such a memory can significantly improve LocAgent, a state-of-the-art localization framework, on both SWE-bench-verified and the more recent SWE-bench-live benchmarks. Our research contributes towards developing agents that can accumulate and leverage past experience for long-horizon tasks, more closely emulating the expertise of human developers.
Authors: Federico Tiblias, Irina Bigoulaeva, Jingcheng Niu, Simone Balloccu, Iryna Gurevych
Abstract: The linear representation hypothesis states that language models (LMs) encode concepts as directions in their latent space, forming organized, multidimensional manifolds. Prior efforts focus on discovering specific geometries for specific features, and thus lack generalization. We introduce Supervised Multi-Dimensional Scaling (SMDS), a model-agnostic method to automatically discover feature manifolds. We apply SMDS to temporal reasoning as a case study, finding that different features form various geometric structures such as circles, lines, and clusters. SMDS reveals many insights on these structures: they consistently reflect the properties of the concepts they represent; are stable across model families and sizes; actively support reasoning in models; and dynamically reshape in response to context changes. Together, our findings shed light on the functional role of feature manifolds, supporting a model of entity-based reasoning in which LMs encode and transform structured representations.
Authors: Xiao Li, Jiaqi Zhang, Shuxiang Zhang, Tianshui Chen, Liang Lin, Guangrun Wang
Abstract: We propose an Authentic Discrete Diffusion (ADD) framework that fundamentally redefines prior pseudo-discrete approaches by preserving core diffusion characteristics directly in the one-hot space through a suite of coordinated mechanisms. Unlike conventional "pseudo" discrete diffusion (PDD) methods, ADD reformulates the diffusion input by directly using float-encoded one-hot class data, without relying on diffusing in the continuous latent spaces or masking policies. At its core, a timestep-conditioned cross-entropy loss is introduced between the diffusion model's outputs and the original one-hot labels. This synergistic design establishes a bridge between discriminative and generative learning. Our experiments demonstrate that ADD not only achieves superior performance on classification tasks compared to the baseline, but also exhibits excellent text generation capabilities on Image captioning. Extensive ablations validate the measurable gains of each component.
Authors: Zichen Liu, Anya Sims, Keyu Duan, Changyu Chen, Simon Yu, Xiangxin Zhou, Haotian Xu, Shaopan Xiong, Bo Liu, Chenmien Tan, Chuen Yang Beh, Weixun Wang, Hao Zhu, Weiyan Shi, Diyi Yang, Michael Shieh, Yee Whye Teh, Wee Sun Lee, Min Lin
Abstract: The training paradigm for large language models (LLMs) is moving from static datasets to experience-based learning, where agents acquire skills via interacting with complex environments. To facilitate this transition we introduce GEM (General Experience Maker), an open-source environment simulator designed for the age of LLMs. Analogous to OpenAI-Gym for traditional reinforcement learning (RL), GEM provides a standardized framework for the environment-agent interface, including asynchronous vectorized execution for high throughput, and flexible wrappers for easy extensibility. GEM also features a diverse suite of environments, robust integrated tools, and single-file example scripts demonstrating using GEM with five popular RL training frameworks. Along with this, we also provide a set of baselines across 24 environments using REINFORCE with Return Batch Normalization (ReBN), which -- unlike GRPO -- is compatible with the full RL setting of dense per-turn rewards and offers better credit assignment. We further conduct apple-to-apple benchmarking of PPO, GRPO and REINFORCE in both single- and multi-turn settings using GEM to shed light on the algorithmic designs. Lastly, GEM also functions as a convenient evaluation toolkit besides a training environment. We hope this framework can help accelerate future agentic LLM research.
Authors: Ruiyi Wang, Prithviraj Ammanabrolu
Abstract: We study what actually works and what doesn't for training large language models as agents via multi-turn reinforcement learning. Despite rapid progress, existing frameworks and definitions are fragmented, and there is no systematic formulation or analysis of which design choices matter across tasks. We address this gap by first breaking down the design space into three inter-related pillars -- environment, reward, and policy -- and empirically derive a recipe for training LLM agents in situated textual domains. In particular, we test TextWorld and ALFWorld, popular domains for testing situated embodied reasoning, as well as SWE-Gym for more software engineering style tasks. (i) For the environment, we analyze the impacts of task complexity in terms of sizes of the state and action spaces as well as optimal solution length, finding that even simple environments within a domain can provide signal on how well an agent can generalize to more complex tasks. (ii) For the reward, we ablate relative reward sparsity, observing that while dense turn-level rewards accelerate training, performance and stability is highly dependent on the choice of RL algorithm. (iii) And for the agent's policy, we explore the interplay between reward sparsity and biased (PPO, GRPO) and unbiased (RLOO) policy gradient methods in addition to showing how to find the optimal Supervised Fine-tuning (SFT) to RL training ratio given a fixed budget. We distill these findings into a training recipe that guides co-design across the three pillars, facilitating research and practical efforts in multi-turn agentic RL. Code: https://github.com/pearls-lab/meow-tea-taro
Authors: Zhaolin Gao, Joongwon Kim, Wen Sun, Thorsten Joachims, Sid Wang, Richard Yuanzhe Pang, Liang Tan
Abstract: We introduce Prompt Curriculum Learning (PCL), a lightweight reinforcement learning (RL) algorithm that selects intermediate-difficulty prompts using a learned value model to post-train language models. Since post-training LLMs via RL remains sensitive to batching and prompt selection strategies, we first conduct a series of systematic experiments where we (1) determine the optimal training batch size that balances generation efficiency and gradient quality and (2) establish the importance of focusing on prompts of intermediate difficulty for the policy. We build upon these results to design PCL, which identifies prompts of intermediate difficulty for the current policy in an on-policy manner by using a value model that is concurrently updated based on the current policy. By focusing on informative prompts that yield high effective ratios, PCL achieves either the highest performance or requires significantly less time to reach comparable performance to its counterparts. Compared to rollout-based filtering methods, PCL avoids costly rollouts and achieves $12.1\times$ and $16.9\times$ faster speed on identifying intermediate-difficulty prompts when training on MATH and DeepScaleR, respectively. We further demonstrate that our value model accurately predicts prompt difficulty and allows PCL to focus on progressively more challenging prompts during RL. Our results present a new methodology that delivers improved tradeoff between upper-bound performance and efficiency for reasoning-focused RL.
Authors: Yiran Shen, Yu Xia, Jonathan Chang, Prithviraj Ammanabrolu
Abstract: Aligning large language models to human preferences is inherently multidimensional, yet most pipelines collapse heterogeneous signals into a single optimizeable objective. We seek to answer what it would take to simultaneously align a model across various domains spanning those with: verifiable rewards (mathematical accuracy), non-verifiable subjective preferences (human values), and complex interactive scenarios (multi-turn AI tutoring dialogues). Such multi-objective reinforcement learning setups are often plagued by the individual objectives being at odds with each other, resulting in inefficient training and little user control during inference. We propose a unified framework that: (i) standardizes {process reward model} (PRM) training across both verifiable and non-verifiable settings to better supervise models' chain-of-thought reasoning; (ii) performs {multi-objective alignment} by training the LLM with our $\textbf{M}$ulti-$\textbf{A}$ction-$\textbf{H}$ead $\textbf{DPO}$ (MAH-DPO) and a vectorized reward where the dimensions of the vector correspond to the various objectives instead of a single scalar; and (iii) demonstrates how such a system provides fine-grained inference-time user control. Experiments across math reasoning, value alignment, and multi-turn dialogue show that our framework improves performance across multiple objectives simultaneously, while minimizing cross-objective trade-offs and enabling flexible inference time user control. The code can be found at https://github.com/pearls-lab/multiobj-align.
Authors: Yanzhe Chen, Kevin Qinghong Lin, Mike Zheng Shou
Abstract: While recent generative models advance pixel-space video synthesis, they remain limited in producing professional educational videos, which demand disciplinary knowledge, precise visual structures, and coherent transitions, limiting their applicability in educational scenarios. Intuitively, such requirements are better addressed through the manipulation of a renderable environment, which can be explicitly controlled via logical commands (e.g., code). In this work, we propose Code2Video, a code-centric agent framework for generating educational videos via executable Python code. The framework comprises three collaborative agents: (i) Planner, which structures lecture content into temporally coherent flows and prepares corresponding visual assets; (ii) Coder, which converts structured instructions into executable Python codes while incorporating scope-guided auto-fix to enhance efficiency; and (iii) Critic, which leverages vision-language models (VLM) with visual anchor prompts to refine spatial layout and ensure clarity. To support systematic evaluation, we build MMMC, a benchmark of professionally produced, discipline-specific educational videos. We evaluate MMMC across diverse dimensions, including VLM-as-a-Judge aesthetic scores, code efficiency, and particularly, TeachQuiz, a novel end-to-end metric that quantifies how well a VLM, after unlearning, can recover knowledge by watching the generated videos. Our results demonstrate the potential of Code2Video as a scalable, interpretable, and controllable approach, achieving 40% improvement over direct code generation and producing videos comparable to human-crafted tutorials. The code and datasets are available at https://github.com/showlab/Code2Video.
Authors: Zhangchen Xu, Adriana Meza Soria, Shawn Tan, Anurag Roy, Ashish Sunil Agrawal, Radha Poovendran, Rameswar Panda
Abstract: Large Language Model (LLM) agents are rapidly emerging as powerful systems for automating tasks across domains. Yet progress in the open-source community is constrained by the lack of high quality permissively licensed tool-agentic training data. Existing datasets are often limited in diversity, realism, and complexity, particularly regarding multi-tool and multi-turn interactions. To address this gap, we introduce Toucan, the largest publicly available tool-agentic dataset to date, containing 1.5 million trajectories synthesized from nearly 500 real-world Model Context Protocols (MCPs). Unlike prior work, Toucan leverages authentic MCP environments to generate diverse, realistic, and challenging tasks with trajectories involving real tool execution. Our pipeline first produces a broad spectrum of tool-use queries using five distinct models, applies model-based quality filtering, and then generates agentic trajectories with three teacher models using two agentic frameworks. Rigorous rule-based and model-based validation ensures high-quality outputs. We also introduce three extension mechanisms to further diversify tasks and simulate multi-turn conversations. Models fine-tuned on Toucan outperform larger closed-source counterparts on the BFCL V3 benchmark and push the Pareto frontier forward on MCP-Universe Bench.
Authors: Jian Hu, Mingjie Liu, Ximing Lu, Fang Wu, Zaid Harchaoui, Shizhe Diao, Yejin Choi, Pavlo Molchanov, Jun Yang, Jan Kautz, Yi Dong
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a key ingredient for unlocking complex reasoning capabilities in large language models. Recent work ProRL has shown promise in scaling RL by increasing the number of training steps. However, performance plateaus after thousands of steps, with clear diminishing returns from allocating more computation to additional training. In this work, we investigate a complementary paradigm for scaling RL, BroR-Lincreasing the number of rollouts per example to hundreds to exhaustively Broaden exploration, which yields continuous performance gains beyond the saturation point observed in ProRL when scaling the number of training steps. Our approach is motivated by a mass balance equation analysis allowing us to characterize the rate of change in probability mass for correct and incorrect tokens during the reinforcement process. We show that under a one-step RL assumption, sampled rollout tokens always contribute to correct-mass expansion, while unsampled tokens outside rollouts may lead to gains or losses depending on their distribution and the net reward balance. Importantly, as the number of rollouts per example N increases, the effect of unsampled terms diminishes, ensuring overall correct-mass expansion. To validate our theoretical analysis, we conduct simulations under more relaxed conditions and find that a sufficiently large rollout size N-corresponding to ample exploration-guarantees an increase in the probability mass of all correct tokens. Empirically, BroRL revives models saturated after 3K ProRL training steps and demonstrates robust, continuous improvement, achieving state-of-the-art results for the 1.5B model across diverse benchmarks.
Authors: Nicolas Yax, Pierre-Yves Oudeyer, Stefano Palminteri
Abstract: This paper introduces PhyloLM, a method adapting phylogenetic algorithms to Large Language Models (LLMs) to explore whether and how they relate to each other and to predict their performance characteristics. Our method calculates a phylogenetic distance metric based on the similarity of LLMs' output. The resulting metric is then used to construct dendrograms, which satisfactorily capture known relationships across a set of 111 open-source and 45 closed models. Furthermore, our phylogenetic distance predicts performance in standard benchmarks, thus demonstrating its functional validity and paving the way for a time and cost-effective estimation of LLM capabilities. To sum up, by translating population genetic concepts to machine learning, we propose and validate a tool to evaluate LLM development, relationships and capabilities, even in the absence of transparent training information.
Authors: Atharvan Dogra, Krishna Pillutla, Ameet Deshpande, Ananya B Sai, John Nay, Tanmay Rajpurohit, Ashwin Kalyan, Balaraman Ravindran
Abstract: We explore the ability of large language models (LLMs) to engage in subtle deception through strategically phrasing and intentionally manipulating information. This harmful behavior can be hard to detect, unlike blatant lying or unintentional hallucination. We build a simple testbed mimicking a legislative environment where a corporate \textit{lobbyist} module is proposing amendments to bills that benefit a specific company while evading identification of this benefactor. We use real-world legislative bills matched with potentially affected companies to ground these interactions. Our results show that LLM lobbyists can draft subtle phrasing to avoid such identification by strong LLM-based detectors. Further optimization of the phrasing using LLM-based re-planning and re-sampling increases deception rates by up to 40 percentage points. Our human evaluations to verify the quality of deceptive generations and their retention of self-serving intent show significant coherence with our automated metrics and also help in identifying certain strategies of deceptive phrasing. This study highlights the risk of LLMs' capabilities for strategic phrasing through seemingly neutral language to attain self-serving goals. This calls for future research to uncover and protect against such subtle deception.
Authors: Jianqing Zhu, Huang Huang, Zhihang Lin, Juhao Liang, Zhengyang Tang, Khalid Almubarak, Abdulmohsen Alharthik, Bang An, Juncai He, Xiangbo Wu, Fei Yu, Junying Chen, Zhuoheng Ma, Yuhao Du, He Zhang, Emad A. Alghamdi, Lian Zhang, Ruoyu Sun, Haizhou Li, Benyou Wang, Jinchao Xu
Abstract: This paper addresses the critical need for democratizing large language models (LLM) in the Arab world, a region that has seen slower progress in developing models comparable to state-of-the-art offerings like GPT-4 or ChatGPT 3.5, due to a predominant focus on mainstream languages (e.g., English and Chinese). One practical objective for an Arabic LLM is to utilize an Arabic-specific vocabulary for the tokenizer that could speed up decoding. However, using a different vocabulary often leads to a degradation of learned knowledge since many words are initially out-of-vocabulary (OOV) when training starts. Inspired by the vocabulary learning during Second Language (Arabic) Acquisition for humans, the released AraLLaMA employs progressive vocabulary expansion, which is implemented by a modified BPE algorithm that progressively extends the Arabic subwords in its dynamic vocabulary during training, thereby balancing the OOV ratio at every stage. The ablation study demonstrated the effectiveness of Progressive Vocabulary Expansion. Moreover, AraLLaMA achieves decent performance comparable to the best Arabic LLMs across a variety of Arabic benchmarks. Models, training data, benchmarks, and codes will be all open-sourced.
Authors: KuanChao Chu, Yi-Pei Chen, Hideki Nakayama
Abstract: Controlling diversity in LLM-agent simulations is essential for balancing stability in structured tasks with variability in open-ended interactions. However, we observe that dialogue diversity tends to degrade over long-term simulations. To explore the role of prompt design in this phenomenon, we modularized the utterance generation prompt and found that reducing contextual information leads to more diverse outputs. Based on this insight, we propose Adaptive Prompt Pruning (APP), a novel method that allows users to control diversity via a single parameter, lambda. APP dynamically prunes prompt segments based on attention scores and is compatible with existing diversity control methods. We demonstrate that APP effectively modulates diversity through extensive experiments and propose a method to balance the control trade-offs. Our analysis reveals that all prompt components impose constraints on diversity, with the Memory being the most influential. Additionally, high-attention contents consistently suppress output diversity.
Authors: Zekun Xi, Wenbiao Yin, Jizhan Fang, Jialong Wu, Runnan Fang, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen, Ningyu Zhang
Abstract: Machine writing with large language models often relies on retrieval-augmented generation. However, these approaches remain confined within the boundaries of the model's predefined scope, limiting the generation of content with rich information. Specifically, vanilla-retrieved information tends to lack depth, novelty, and suffers from redundancy, which negatively impacts the quality of generated articles, leading to shallow, unoriginal, and repetitive outputs. To address these issues, we propose OmniThink, a slow-thinking machine writing framework that emulates the human-like process of iterative expansion and reflection. The core idea behind OmniThink is to simulate the cognitive behavior of learners as they slowly deepen their knowledge of the topics. Experimental results demonstrate that OmniThink improves the knowledge density of generated articles without compromising metrics such as coherence and depth. Human evaluations and expert feedback further highlight the potential of OmniThink to address real-world challenges in the generation of long-form articles. Code is available at https://github.com/zjunlp/OmniThink.
Authors: Xiaoyang Liu, Kangjie Bao, Jiashuo Zhang, Yunqi Liu, Yu Chen, Yuntian Liu, Yang Jiao, Tao Luo
Abstract: Autoformalization, the automatic translation of mathematical content from natural language into machine-verifiable formal languages, has seen significant progress driven by advances in large language models (LLMs). Nonetheless, a primary barrier to further improvements is the limited availability of parallel corpora that map informal mathematical text to its formal counterpart. To address this limitation, we propose ATLAS (Autoformalizing Theorems through Lifting, Augmentation, and Synthesis of Data), a novel data generation framework designed to produce large-scale, high-quality parallel corpora of theorem statements. Distinct from prior approaches, ATLAS begins with a concept repository, accelerates the improvement of the student model through expert iteration combined with knowledge distillation, and introduces two novel augmentation strategies that exploit the structural characteristics of formal languages. Running the proposed ATLAS framework for 10 iterations, we construct an undergraduate-level dataset of 117k theorem statements and develop the ATLAS Translator by fine-tuning Llama3.1-8B-Instruct with LoRA. This model establishes a new state of the art, demonstrating statistically significant improvements over both the Herald Translator and the Kimina-Autoformalizer across all benchmarks (p<0.05, two-sided t-test). Furthermore, we demonstrate that the full-parameter fine-tuning of a stronger base model on the ATLAS dataset leads to superior performance. The datasets, model, and code are available at https://github.com/XiaoyangLiu-sjtu/ATLAS.
Authors: Bhiman Kumar Baghel, Emma Jordan, Zheyuan Ryan Shi, Xiang Lorraine Li
Abstract: Large Language Models (LLMs) are widely deployed in downstream tasks, but keeping their knowledge up-to-date via retraining or fine-tuning is often computationally expensive. Model editing provides a more efficient alternative by updating a targeted subset of parameters, which often follows the locate-and-edit paradigm. Despite this efficiency, existing methods are limited: edits may fail to inject knowledge (UnderEdit) or unintentionally disrupt unrelated neighboring knowledge (OverEdit). To address these challenges, we propose two complementary methods: iterative model editing, which applies successive edits to mitigate UnderEdit, and neighbor-assisted model editing, which incorporates neighboring knowledge during editing to reduce OverEdit. Our extensive experiments show that these techniques improve editing performance across multiple LLMs, algorithms, and benchmarks, reducing UnderEdit by up to 38 percentage points and OverEdit by up to 6, while remaining broadly applicable to any locate-and-edit method. We release our code at https://github.com/bhimanbaghel/ResolveUnderOverEdit.
Authors: Mohammad Almansoori, Komal Kumar, Hisham Cholakkal
Abstract: In this work, we introduce MedAgentSim, an open-source simulated clinical environment with doctor, patient, and measurement agents designed to evaluate and enhance LLM performance in dynamic diagnostic settings. Unlike prior approaches, our framework requires doctor agents to actively engage with patients through multi-turn conversations, requesting relevant medical examinations (e.g., temperature, blood pressure, ECG) and imaging results (e.g., MRI, X-ray) from a measurement agent to mimic the real-world diagnostic process. Additionally, we incorporate self improvement mechanisms that allow models to iteratively refine their diagnostic strategies. We enhance LLM performance in our simulated setting by integrating multi-agent discussions, chain-of-thought reasoning, and experience-based knowledge retrieval, facilitating progressive learning as doctor agents interact with more patients. We also introduce an evaluation benchmark for assessing the LLM's ability to engage in dynamic, context-aware diagnostic interactions. While MedAgentSim is fully automated, it also supports a user-controlled mode, enabling human interaction with either the doctor or patient agent. Comprehensive evaluations in various simulated diagnostic scenarios demonstrate the effectiveness of our approach. Our code, simulation tool, and benchmark are available at \href{https://medagentsim.netlify.app/}.
Authors: Maxime Bouthors, Josep Crego, Fran\c{c}ois Yvon
Abstract: Conventional retrieval-augmented neural machine translation (RANMT) systems leverage bilingual corpora, e.g., translation memories (TMs). Yet, in many settings, monolingual corpora in the target language are often available. This work explores ways to take advantage of such resources by directly retrieving relevant target language segments, based on a source-side query. For this, we design improved cross-lingual retrieval systems, trained with both sentence level and word-level matching objectives. In our experiments with three RANMT architectures, we assess such cross-lingual objectives in a controlled setting, reaching performances that match those of standard TM-based models. We also showcase our method on a real-world settings, using much larger monolingual and observe strong improvements over both the baseline setting and general-purpose cross-lingual retrievers.
Authors: Zhibo Hu, Chen Wang, Yanfeng Shu, Hye-Young Paik, Liming Zhu
Abstract: Ambiguity in natural language is a significant obstacle for achieving accurate text to structured data mapping through large language models (LLMs), which affects the performance of tasks such as mapping text to agentic tool calling and text-to-SQL queries. Existing methods to ambiguity handling either rely on the ReACT framework to obtain correct mappings through trial and error, or on supervised fine-tuning to bias models toward specific tasks. In this paper, we adopt a different approach that characterizes representation differences of ambiguous text in the latent space and leverages these differences to identify ambiguity before mapping them to structured data. To detect sentence-level ambiguity, we focus on the relationship between ambiguous questions and their interpretations. Unlike distances calculated by dense embeddings, we introduce a new distance measure based on a path kernel over concepts. With this measurement, we identify patterns to distinguish ambiguous from unambiguous questions. Furthermore, we propose a method for improving LLM performance on ambiguous agentic tool calling through missing concept prediction. Both achieve state-of-the-art results.
Authors: Daehee Kim, Deokhyung Kang, Jonghwi Kim, Sangwon Ryu, Gary Geunbae Lee
Abstract: Legal Passage Retrieval (LPR) systems are crucial as they help practitioners save time when drafting legal arguments. However, it remains an underexplored avenue. One primary reason is the significant vocabulary mismatch between the query and the target passage. To address this, we propose a simple yet effective method, the Generative query REwriter (GuRE). We leverage the generative capabilities of Large Language Models (LLMs) by training the LLM for query rewriting. "Rewritten queries" help retrievers to retrieve target passages by mitigating vocabulary mismatch. Experimental results show that GuRE significantly improves performance in a retriever-agnostic manner, outperforming all baseline methods. Further analysis reveals that different training objectives lead to distinct retrieval behaviors, making GuRE more suitable than direct retriever fine-tuning for real-world applications. Codes are avaiable at github.com/daehuikim/GuRE.
Authors: Joakim Edin, R\'obert Csord\'as, Tuukka Ruotsalo, Zhengxuan Wu, Maria Maistro, Casper L. Christensen, Jing Huang, Lars Maal{\o}e
Abstract: Ensuring faithful interpretability in large language models is imperative for trustworthy and reliable AI. A key obstacle is self-repair, a phenomenon where networks compensate for reduced signal in one component by amplifying others, masking the true importance of the ablated component. While prior work attributes self-repair to layer normalization and back-up components that compensate for ablated components, we identify a novel form occurring within the attention mechanism, where softmax redistribution conceals the influence of important attention scores. This leads traditional ablation and gradient-based methods to underestimate the significance of all components contributing to these attention scores. We introduce Gradient Interaction Modifications (GIM), a technique that accounts for self-repair during backpropagation. Extensive experiments across multiple large language models (Gemma 2B/9B, LLAMA 1B/3B/8B, Qwen 1.5B/3B) and diverse tasks demonstrate that GIM significantly improves faithfulness over existing circuit identification and feature attribution methods. Our work is a significant step toward better understanding the inner mechanisms of LLMs, which is crucial for improving them and ensuring their safety. Our code is available at https://github.com/JoakimEdin/gim.
Authors: Jiwan Chung, Junhyeok Kim, Siyeol Kim, Jaeyoung Lee, Min Soo Kim, Youngjae Yu
Abstract: When thinking with images, humans rarely rely on a single glance: they revisit visual information repeatedly during reasoning. However, existing models typically process images only once and thereafter generate reasoning entirely in text, lacking mechanisms to re-access or ground inference in visual representations. We empirically confirm this: as reasoning chains lengthen, models progressively lose focus on relevant regions. In response, we introduce v1, a lightweight extension that enables active visual referencing through a simple point-and-copy approach. This allows the model to identify relevant image patches and copy their embeddings back into the reasoning stream, ensuring that evolving hypotheses remain grounded in perceptual evidence. Crucially, our pointing strategy lets the MLLM directly select image patches using their semantic representations as keys, keeping perceptual evidence embedded in the same space as the model's reasoning. To train this capability, we construct v1g, a dataset of 300K multimodal reasoning traces with interleaved visual grounding annotations. Across various multimodal mathematical reasoning benchmarks, v1 consistently outperforms comparable baselines, establishing point-and-copy as a practical mechanism for grounded reasoning. The model checkpoint and dataset are available at github.com/jun297/v1.
Authors: Wesley Scivetti, Tatsuya Aoyama, Ethan Wilcox, Nathan Schneider
Abstract: Humans have a remarkable ability to acquire and understand grammatical phenomena that are seen rarely, if ever, during childhood. Recent evidence suggests that language models with human-scale pretraining data may possess a similar ability by generalizing from frequent to rare constructions. However, it remains an open question how widespread this generalization ability is, and to what extent this knowledge extends to meanings of rare constructions, as opposed to just their forms. We fill this gap by testing human-scale transformer language models on their knowledge of both the form and meaning of the (rare and quirky) English LET-ALONE construction. To evaluate our LMs we construct a bespoke synthetic benchmark that targets syntactic and semantic properties of the construction. We find that human-scale LMs are sensitive to form, even when related constructions are filtered from the dataset. However, human-scale LMs do not make correct generalizations about LET-ALONE's meaning. These results point to an asymmetry in the current architectures' sample efficiency between language form and meaning, something which is not present in human language learners.
Authors: Hongbo Zhao, Fei Zhu, Haiyang Guo, Meng Wang, Rundong Wang, Gaofeng Meng, Zhaoxiang Zhang
Abstract: Recent Multimodal Large Language Models (MLLMs) excel in vision-language understanding but face challenges in adapting to dynamic real-world scenarios that require continuous integration of new knowledge and skills. While continual learning (CL) offers a potential solution, existing benchmarks and methods suffer from critical limitations. In this paper, we introduce MLLM-CL, a novel benchmark encompassing domain and ability continual learning, where the former focuses on independently and identically distributed (IID) evaluation across evolving mainstream domains, whereas the latter evaluates on non-IID scenarios with new model abilities. Methodologically, we propose preventing catastrophic interference through parameter isolation and an MLLM-based routing mechanism. Extensive experiments demonstrate that our approach can integrate domain-specific knowledge and functional abilities with minimal forgetting, significantly outperforming existing methods. Our benchmark and code are available at https://github.com/bjzhb666/MLLM-CL.
Authors: Jacqueline He, Howard Yen, Margaret Li, Shuyue Stella Li, Zhiyuan Zeng, Weijia Shi, Yulia Tsvetkov, Danqi Chen, Pang Wei Koh, Luke Zettlemoyer
Abstract: A central challenge in language models (LMs) is faithfulness hallucination: the generation of information unsubstantiated by input context. To study this problem, we propose Precise Information Control (PIC), a new task formulation that requires models to generate long-form outputs grounded in a provided set of short self-contained statements, without adding any unsupported ones. PIC includes a full setting that tests a model's ability to include exactly all input claims, and a partial setting that requires the model to selectively incorporate only relevant claims. We present PIC-Bench, a benchmark of eight long-form generation tasks (e.g., summarization, biography generation) adapted to the PIC setting, where LMs are supplied with well-formed, verifiable input claims. Our evaluation of a range of open and proprietary LMs on PIC-Bench reveals that, surprisingly, state-of-the-art LMs still hallucinate against user-provided input in over 70% of generations. To alleviate this lack of faithfulness, we introduce a post-training framework that uses a weakly supervised preference data construction method to train an 8B PIC-LM with stronger PIC ability--improving from 69.1% to 91.0% F1 in the full PIC setting. When integrated into end-to-end factual generation pipelines, PIC-LM improves exact match recall by 17.1% on ambiguous QA with retrieval, and factual precision by 30.5% on a birthplace fact-checking task, underscoring the potential of precisely grounded generation.
Authors: Renjie Luo, Jiaxi Li, Chen Huang, Wei Lu
Abstract: Long chain-of-thought (CoT) supervision has become a common strategy to enhance reasoning in language models. While effective for large models, we identify a phenomenon we call Long CoT Degradation, in which small language models (SLMs; <=3B parameters) trained on limited long CoT data experience significant performance deterioration. Through extensive experiments on the Qwen2.5, LLaMA3 and Gemma3 families, we demonstrate that this degradation is widespread across SLMs. In some settings, models trained on only 8k long CoT examples lose up to 75% of their original performance before fine-tuning. Strikingly, we further observe that for some particularly small models, even training on 220k long CoT examples fails to recover or surpass their original performance prior to fine-tuning. Our analysis attributes this effect to error accumulation: while longer responses increase the capacity for multi-step reasoning, they also amplify the risk of compounding mistakes. Furthermore, we find that Long CoT Degradation may negatively impacts downstream reinforcement learning (RL), although this can be alleviated by sufficiently scaled supervised fine-tuning (SFT). Our findings challenge common assumptions about the benefits of long CoT training for SLMs and offer practical guidance for building more effective small-scale reasoning models.
Authors: Li-Ming Zhan, Bo Liu, Chengqiang Xie, Jiannong Cao, Xiao-Ming Wu
Abstract: Inference-time steering aims to alter a large language model's (LLM's) responses without changing its parameters, but a central challenge is identifying the internal modules that most strongly govern the target behavior. Existing approaches often rely on simplistic cues or ad hoc heuristics, leading to suboptimal or unintended effects. We introduce REAL, a framework for identifying behavior-relevant modules (attention heads or layers) in Transformer models. For each module, REAL trains a vector-quantized autoencoder (VQ-AE) on its hidden activations and uses a shared, learnable codebook to partition the latent space into behavior-relevant and behavior-irrelevant subspaces. REAL quantifies a module's behavioral relevance by how well its VQ-AE encodings discriminate behavior-aligned from behavior-violating responses via a binary classification metric; this score guides both module selection and steering strength. We evaluate REAL across eight LLMs from the Llama and Qwen families and nine datasets spanning truthfulness enhancement, open-domain QA under knowledge conflicts, and general alignment tasks. REAL enables more effective inference-time interventions, achieving an average relative improvement of 20% (up to 81.5%) over the ITI method on truthfulness steering. In addition, the modules selected by REAL exhibit strong zero-shot generalization in cross-domain truthfulness-steering scenarios.
Authors: Yahan Li, Jifan Yao, John Bosco S. Bunyi, Adam C. Frank, Angel Hwang, Ruishan Liu
Abstract: Medical question answering (QA) benchmarks often focus on multiple-choice or fact-based tasks, leaving open-ended answers to real patient questions underexplored. This gap is particularly critical in mental health, where patient questions often mix symptoms, treatment concerns, and emotional needs, requiring answers that balance clinical caution with contextual sensitivity. We present CounselBench, a large-scale benchmark developed with 100 mental health professionals to evaluate and stress-test large language models (LLMs) in realistic help-seeking scenarios. The first component, CounselBench-EVAL, contains 2,000 expert evaluations of answers from GPT-4, LLaMA 3, Gemini, and human therapists on patient questions from the public forum CounselChat. Each answer is rated across six clinically grounded dimensions, with span-level annotations and written rationales. Expert evaluations show that while LLMs achieve high scores on several dimensions, they also exhibit recurring issues, including unconstructive feedback, overgeneralization, and limited personalization or relevance. Responses were frequently flagged for safety risks, most notably unauthorized medical advice. Follow-up experiments show that LLM judges systematically overrate model responses and overlook safety concerns identified by human experts. To probe failure modes more directly, we construct CounselBench-Adv, an adversarial dataset of 120 expert-authored mental health questions designed to trigger specific model issues. Evaluation of 3,240 responses from nine LLMs reveals consistent, model-specific failure patterns. Together, CounselBench establishes a clinically grounded framework for benchmarking LLMs in mental health QA.
Authors: Tommaso Green, Martin Gubri, Haritz Puerto, Sangdoo Yun, Seong Joon Oh
Abstract: We study privacy leakage in the reasoning traces of large reasoning models used as personal agents. Unlike final outputs, reasoning traces are often assumed to be internal and safe. We challenge this assumption by showing that reasoning traces frequently contain sensitive user data, which can be extracted via prompt injections or accidentally leak into outputs. Through probing and agentic evaluations, we demonstrate that test-time compute approaches, particularly increased reasoning steps, amplify such leakage. While increasing the budget of those test-time compute approaches makes models more cautious in their final answers, it also leads them to reason more verbosely and leak more in their own thinking. This reveals a core tension: reasoning improves utility but enlarges the privacy attack surface. We argue that safety efforts must extend to the model's internal thinking, not just its outputs.
Authors: Matteo Guida, Yulia Otmakhova, Eduard Hovy, Lea Frermann
Abstract: When a news article describes immigration as an "economic burden" or a "humanitarian crisis," it selectively emphasizes certain aspects of the issue. Although \textit{framing} shapes how the public interprets such issues, audiences do not absorb frames passively but actively reorganize the presented information. While this relationship between source content and audience response is well-documented in the social sciences, NLP approaches often ignore it, detecting frames in articles and responses in isolation. We present the first computational framework for large-scale analysis of framing across source content (news articles) and audience responses (reader comments). Methodologically, we refine frame labels and develop a framework that reconstructs dominant frames in articles and comments from sentence-level predictions, and aligns articles with topically relevant comments. Applying our framework across eleven topics and two news outlets, we find that frame reuse in comments correlates highly across outlets, while topic-specific patterns vary. We release a frame classifier that performs well on both articles and comments, a dataset of article and comment sentences manually labeled for frames, and a large-scale dataset of articles and comments with predicted frame labels.
Authors: Xi Ai, Mahardika Krisna Ihsani, Min-Yen Kan
Abstract: Cross-lingual consistency should be considered to assess cross-lingual transferability, maintain the factuality of the model knowledge across languages, and preserve the parity of language model performance. We are thus interested in analyzing, evaluating, and interpreting cross-lingual consistency for factual knowledge. To facilitate our study, we examine multiple pretrained models and tuned models with code-mixed coreferential statements that convey identical knowledge across languages. Interpretability approaches are leveraged to analyze the behavior of a model in cross-lingual contexts, showing different levels of consistency in multilingual models, subject to language families, linguistic factors, scripts, and a bottleneck in cross-lingual consistency on a particular layer. Code-switching training and cross-lingual word alignment objectives show the most promising results, emphasizing the worthiness of cross-lingual alignment supervision and code-switching strategies for both multilingual performance and cross-lingual consistency enhancement. In addition, experimental results suggest promising result for calibrating consistency in the test time via activation patching.
Authors: Mohammad Aflah Khan, Ameya Godbole, Johnny Tian-Zheng Wei, Ryan Wang, James Flemings, Krishna P. Gummadi, Willie Neiswanger, Robin Jia
Abstract: Understanding the relationship between training data and model behavior during pretraining is crucial, but existing workflows make this process cumbersome, fragmented, and often inaccessible to researchers. We present TokenSmith, an open-source library for interactive editing, inspection, and analysis of datasets used in Megatron-style pretraining frameworks such as GPT-NeoX, Megatron, and NVIDIA NeMo. TokenSmith supports a wide range of operations including searching, viewing, ingesting, exporting, inspecting, and sampling data, all accessible through a simple user interface and a modular backend. It also enables structured editing of pretraining data without requiring changes to training code, simplifying dataset debugging, validation, and experimentation. TokenSmith is designed as a plug-and-play addition to existing large language model pretraining workflows, thereby democratizing access to production-grade dataset tooling. TokenSmith is hosted on GitHub, with accompanying documentation, tutorials, and a demonstration video (available on YouTube).
Authors: Muhammed Saeed, Shaina Raza, Ashmal Vayani, Muhammad Abdul-Mageed, Ali Emami, Shady Shehata
Abstract: Research on bias in Text-to-Image (T2I) models has primarily focused on demographic representation and stereotypical attributes, overlooking a fundamental question: how does grammatical gender influence visual representation across languages? We introduce a cross-linguistic benchmark examining words where grammatical gender contradicts stereotypical gender associations (e.g., ``une sentinelle'' - grammatically feminine in French but referring to the stereotypically masculine concept ``guard''). Our dataset spans five gendered languages (French, Spanish, German, Italian, Russian) and two gender-neutral control languages (English, Chinese), comprising 800 unique prompts that generated 28,800 images across three state-of-the-art T2I models. Our analysis reveals that grammatical gender dramatically influences image generation: masculine grammatical markers increase male representation to 73% on average (compared to 22% with gender-neutral English), while feminine grammatical markers increase female representation to 38% (compared to 28% in English). These effects vary systematically by language resource availability and model architecture, with high-resource languages showing stronger effects. Our findings establish that language structure itself, not just content, shapes AI-generated visual outputs, introducing a new dimension for understanding bias and fairness in multilingual, multimodal systems.
Authors: Bram van Dijk, Tiberon Kuiper, Sirin Aoulad si Ahmed, Armel Levebvre, Jake Johnson, Jan Duin, Simon Mooijaart, Marco Spruit
Abstract: Voice-controlled interfaces can support older adults in clinical contexts -- with chatbots being a prime example -- but reliable Automatic Speech Recognition (ASR) for underrepresented groups remains a bottleneck. This study evaluates state-of-the-art ASR models on language use of older Dutch adults, who interacted with the Welzijn.AI chatbot designed for geriatric contexts. We benchmark generic multilingual ASR models, and models fine-tuned for Dutch spoken by older adults, while also considering processing speed. Our results show that generic multilingual models outperform fine-tuned models, which suggests recent ASR models can generalise well out of the box to real-world datasets. Moreover, our results indicate that truncating generic models is helpful in balancing the accuracy-speed trade-off. Nonetheless, we also find inputs which cause a high word error rate and place them in context.
Authors: Hunzalah Hassan Bhatti, Youssef Ahmed, Md Arid Hasan, Firoj Alam
Abstract: In this paper, we report our participation to the PalmX cultural evaluation shared task. Our system, CultranAI, focused on data augmentation and LoRA fine-tuning of large language models (LLMs) for Arabic cultural knowledge representation. We benchmarked several LLMs to identify the best-performing model for the task. In addition to utilizing the PalmX dataset, we augmented it by incorporating the Palm dataset and curated a new dataset of over 22K culturally grounded multiple-choice questions (MCQs). Our experiments showed that the Fanar-1-9B-Instruct model achieved the highest performance. We fine-tuned this model on the combined augmented dataset of 22K+ MCQs. On the blind test set, our submitted system ranked 5th with an accuracy of 70.50%, while on the PalmX development set, it achieved an accuracy of 84.1%.
Authors: Zifeng Cheng, Jinwei Gan, Zhiwei Jiang, Cong Wang, Yafeng Yin, Xiang Luo, Yuchen Fu, Qing Gu
Abstract: Large language models (LLMs) have achieved remarkable performance across many generation tasks. Nevertheless, effectively aligning them with desired behaviors remains a significant challenge. Activation steering is an effective and cost-efficient approach that directly modifies the activations of LLMs during the inference stage, aligning their responses with the desired behaviors and avoiding the high cost of fine-tuning. Existing methods typically indiscriminately intervene to all generations or rely solely on the question to determine intervention, which limits the accurate assessment of the intervention strength. To this end, we propose the Flexible Activation Steering with Backtracking (FASB) framework, which dynamically determines both the necessity and strength of intervention by tracking the internal states of the LLMs during generation, considering both the question and the generated content. Since intervening after detecting a deviation from the desired behavior is often too late, we further propose the backtracking mechanism to correct the deviated tokens and steer the LLMs toward the desired behavior. Extensive experiments on the TruthfulQA dataset and six multiple-choice datasets demonstrate that our method outperforms baselines. Our code will be released at https://github.com/gjw185/FASB.
Authors: Seiji Maekawa, Hayate Iso, Nikita Bhutani
Abstract: Effective decision-making often relies on identifying what makes each candidate distinctive. While existing benchmarks for LLMs emphasize retrieving or summarizing information relevant to a given query, they do not evaluate a model's ability to identify globally distinctive features across a set of documents. We introduce Distinctive Feature Mining (DFM), a new task that challenges models to analyze a small-to-medium collection (10-40 documents) and surface features that are rare in the global context (e.g., appearing in less than 10% of documents). This setting mirrors real-world scenarios such as candidate selection or product differentiation, where statistical reasoning, not retrieval, is key. To enable systematic evaluation of this capability, we present DiFBench, a configurable benchmark creation framework with controllable parameters such as document set size and distinctiveness thresholds. Using DiFBench, we perform a large-scale assessment of distinctive feature mining across ten state-of-the-art LLMs. Our findings reveal a significant performance gap between general-purpose and reasoning-enhanced models. All models, however, substantially degrade as the task complexity and document count increase. We also find that a common failure mode is misidentifying frequent features as distinctive. These insights reveal core limitations in contemporary LLMs' abilities to perform fine-grained, statistical reasoning and rarity detection.
Authors: Sandro Tsang
Abstract: Background: Evidence synthesis facilitates evidence-based medicine. This task becomes increasingly difficult to accomplished with applying computational solutions, since the medical literature grows at astonishing rates. Objective: This study evaluates an information retrieval-driven workflow, CASMA, to enhance the efficiency, transparency, and reproducibility of systematic reviews. Endometriosis recurrence serves as the ideal case due to its complex and ambiguous literature. Methods: The hybrid approach integrates PRISMA guidelines with fuzzy matching and regular expression (regex) to facilitate semi-automated deduplication and filtered records before manual screening. The workflow synthesised evidence from randomised controlled trials on the efficacy of a subclass of gonadotropin-releasing hormone agonists (GnRH-a). A modified splitting method addressed unit-of-analysis errors in multi-arm trials. Results: The workflow sharply reduced the screening workload, taking only 11 days to fetch and filter 33,444 records. Seven eligible RCTs were synthesized (841 patients). The pooled random-effects model yielded a Risk Ratio (RR) of $0.64$ ($95\%$ CI $0.48$ to $0.86$), demonstrating a $36\%$ reduction in recurrence, with non-significant heterogeneity ($I^2=0.00\%$, $\tau^2=0.00$). The findings were robust and stable, as they were backed by sensitivity analyses. Conclusion: This study demonstrates an application of an information-retrieval-driven workflow for medical evidence synthesis. The approach yields valuable clinical results and a generalisable framework to scale up the evidence synthesis, bridging the gap between clinical research and computer science.
Authors: Sujeong Lee, Hayoung Lee, Seongsoo Heo, Wonik Choi
Abstract: Reliable evaluation of large language models is essential to ensure their applicability in practical scenarios. Traditional benchmark-based evaluation methods often rely on fixed reference answers, limiting their ability to capture important qualitative aspects of generated responses. To address these shortcomings, we propose an integrated evaluation framework called \textit{self-refining descriptive evaluation with expert-driven diagnostics}, SPEED, which utilizes specialized functional experts to perform comprehensive, descriptive analyses of model outputs. Unlike conventional approaches, SPEED actively incorporates expert feedback across multiple dimensions, including hallucination detection, toxicity assessment, and lexical-contextual appropriateness. Experimental results demonstrate that SPEED achieves robust and consistent evaluation performance across diverse domains and datasets. Additionally, by employing relatively compact expert models, SPEED demonstrates superior resource efficiency compared to larger-scale evaluators. These findings illustrate that SPEED significantly enhances fairness and interpretability in LLM evaluations, offering a promising alternative to existing evaluation methodologies.
Authors: Jean Seo, Hyunkyung Lee, Gibaeg Kim, Wooseok Han, Jaehyo Yoo, Seungseop Lim, Kihun Shin, Eunho Yang
Abstract: Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods--such as guardrails and tool calling--often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS (TAxonomy of COmprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS is a taxonomy that can cover a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal useful insights about train data distribution and pretrained knowledge of base models.
Authors: Pengfei Wang, Baolin Sun, Xuemei Dong, Yaxun Dai, Hongwei Yuan, Mengdie Chu, Yingqi Gao, Xiang Qi, Peng Zhang, Ying Yan
Abstract: State-of-the-art (SOTA) Text-to-SQL methods still lag significantly behind human experts on challenging benchmarks like BIRD. Current approaches that explore test-time scaling lack an orchestrated strategy and neglect the model's internal reasoning process. To bridge this gap, we introduce Agentar-Scale-SQL, a novel framework leveraging scalable computation to improve performance. Agentar-Scale-SQL implements an Orchestrated Test-Time Scaling strategy that synergistically combines three distinct perspectives: i) Internal Scaling via RL-enhanced Intrinsic Reasoning, ii) Sequential Scaling through Iterative Refinement, and iii) Parallel Scaling using Diverse Synthesis and Tournament Selection. Agentar-Scale-SQL is a general-purpose framework designed for easy adaptation to new databases and more powerful language models. Extensive experiments show that Agentar-Scale-SQL achieves SOTA performance on the BIRD benchmark, reaching 81.67% execution accuracy on the test set and ranking first on the official leaderboard, demonstrating an effective path toward human-level performance.
Authors: Matteo Fuoli, Weihang Huang, Jeannette Littlemore, Sarah Turner, Ellen Wilding
Abstract: Metaphor is a pervasive feature of discourse and a powerful lens for examining cognition, emotion, and ideology. Large-scale analysis, however, has been constrained by the need for manual annotation due to the context-sensitive nature of metaphor. This study investigates the potential of large language models (LLMs) to automate metaphor identification in full texts. We compare three methods: (i) retrieval-augmented generation (RAG), where the model is provided with a codebook and instructed to annotate texts based on its rules and examples; (ii) prompt engineering, where we design task-specific verbal instructions; and (iii) fine-tuning, where the model is trained on hand-coded texts to optimize performance. Within prompt engineering, we test zero-shot, few-shot, and chain-of-thought strategies. Our results show that state-of-the-art closed-source LLMs can achieve high accuracy, with fine-tuning yielding a median F1 score of 0.79. A comparison of human and LLM outputs reveals that most discrepancies are systematic, reflecting well-known grey areas and conceptual challenges in metaphor theory. We propose that LLMs can be used to at least partly automate metaphor identification and can serve as a testbed for developing and refining metaphor identification protocols and the theory that underpins them.
Authors: Changsheng Zhao, Ernie Chang, Zechun Liu, Chia-Jung Chang, Wei Wen, Chen Lai, Sheng Cao, Yuandong Tian, Raghuraman Krishnamoorthi, Yangyang Shi, Vikas Chandra
Abstract: The paradigm shift in large language models (LLMs) from instinctive responses to chain-of-thought (CoT) reasoning has fueled two prevailing assumptions: (1) reasoning capabilities only emerge in sufficiently large models, and (2) such capabilities require training on massive datasets. While the first assumption has already been challenged by recent sub-billion-parameter reasoning models such as Qwen3-0.6B and DeepSeek distilled variants, the second remains largely unquestioned. In this work, we revisit the necessity of scaling to extremely large corpora (>10T tokens) for reasoning emergence. By carefully curating and resampling open-source datasets that we identify as beneficial under our designed metrics, we demonstrate that strong reasoning abilities can emerge with far less data. Specifically, we show that only ~2T tokens of high-quality data are sufficient, and pre-training with 4.2T tokens on the dataset resampled from these ~2T tokens, followed by a established post-training procedure, enables the development of MobileLLM-R1, a series of sub-billion-parameter reasoning models that substantially outperform prior models trained on fully open-sourced data. For example, MobileLLM-R1-950M achieves an AIME score of 15.5, compared to just 0.6 for OLMo-2-1.48B and 0.3 for SmolLM-2-1.7B. Remarkably, despite being trained on only 11.7% of the tokens compared to Qwen3's proprietary 36T-token corpus for pretraining, MobileLLM-R1-950M matches or surpasses Qwen3-0.6B across multiple reasoning benchmarks. To facilitate further research in this direction, we have released the complete training recipe, data sources, data mixing ratio, and model checkpoints, together with the key insights obtained throughout this study.
Authors: Haoyang Zheng, Xinyang Liu, Cindy Xiangrui Kong, Nan Jiang, Zheyuan Hu, Weijian Luo, Wei Deng, Guang Lin
Abstract: Fast and high-quality language generation is the holy grail that people pursue in the age of AI. In this work, we introduce Discrete Diffusion Divergence Instruct (DiDi-Instruct), a training-based method that initializes from a pre-trained (masked) discrete diffusion language model (dLLM) and distills a few-step student for fast generation. The resulting DiDi-Instruct model achieves comparable or superior performance to its dLLM teacher and the GPT-2 baseline while enabling up to 64$\times$ acceleration. The theoretical foundation of DiDi-Instruct is a novel framework based on integral KL-divergence minimization, which yields a practical training algorithm. We further introduce grouped reward normalization, intermediate-state matching, and the reward-guided ancestral sampler that significantly improve training stability, model coverage, and inference quality. On OpenWebText, DiDi-Instruct achieves perplexity from 62.2 (8 NFEs) to 18.4 (128 NFEs), which outperforms prior accelerated dLLMs and GPT-2 baseline. These gains come with a negligible entropy loss (around $1\%$) and reduce additional training wall-clock time by more than $20\times$ compared to competing dLLM distillation methods. We further validate the robustness and effectiveness of DiDi-Instruct through extensive ablation studies, model scaling, and the generation of discrete protein sequences. In conclusion, DiDi-Instruct is an efficient yet effective distillation method, enabling language generation in the blink of an eye. We will release both code and models at github.com/haoyangzheng-ai/didi-instruct.
Authors: Feng Wang, Yuqing Li, Han Xiao
Abstract: jina-reranker-v3 is a 0.6B parameter multilingual document reranker that introduces a novel last but not late interaction. Unlike late interaction models such as ColBERT that perform separate encoding followed by multi-vector matching, our approach conducts causal self-attention between query and documents within the same context window, enabling rich cross-document interactions before extracting contextual embeddings from the last token of each document. This compact architecture achieves state-of-the-art BEIR performance with 61.94 nDCG@10 while being significant smaller than generative listwise rerankers.
Authors: Penghai Zhao, Jinyu Tian, Qinghua Xing, Xin Zhang, Zheng Li, Jianjun Qian, Ming-Ming Cheng, Xiang Li
Abstract: The ability to estimate the quality of scientific papers is central to how both humans and AI systems will advance scientific knowledge in the future. However, existing LLM-based estimation methods suffer from high inference cost, whereas the faster direct score regression approach is limited by scale inconsistencies. We present NAIPv2, a debiased and efficient framework for paper quality estimation. NAIPv2 employs pairwise learning within domain-year groups to reduce inconsistencies in reviewer ratings and introduces the Review Tendency Signal (RTS) as a probabilistic integration of reviewer scores and confidences. To support training and evaluation, we further construct NAIDv2, a large-scale dataset of 24,276 ICLR submissions enriched with metadata and detailed structured content. Trained on pairwise comparisons but enabling efficient pointwise prediction at deployment, NAIPv2 achieves state-of-the-art performance (78.2% AUC, 0.432 Spearman), while maintaining scalable, linear-time efficiency at inference. Notably, on unseen NeurIPS submissions, it further demonstrates strong generalization, with predicted scores increasing consistently across decision categories from Rejected to Oral. These findings establish NAIPv2 as a debiased and scalable framework for automated paper quality estimation, marking a step toward future scientific intelligence systems. Code and dataset are released at sway.cloud.microsoft/Pr42npP80MfPhvj8.
Authors: Tadesse Destaw Belay, Kedir Yassin Hussen, Sukairaj Hafiz Imam, Ibrahim Said Ahmad, Isa Inuwa-Dutse, Abrham Belete Haile, Grigori Sidorov, Iqra Ameer, Idris Abdulmumin, Tajuddeen Gwadabe, Vukosi Marivate, Seid Muhie Yimam, Shamsuddeen Hassan Muhammad
Abstract: Natural Language Processing (NLP) is undergoing constant transformation, as Large Language Models (LLMs) are driving daily breakthroughs in research and practice. In this regard, tracking the progress of NLP research and automatically analyzing the contributions of research papers provides key insights into the nature of the field and the researchers. This study explores the progress of African NLP (AfricaNLP) by asking (and answering) basic research questions such as: i) How has the nature of NLP evolved over the last two decades?, ii) What are the contributions of AfricaNLP papers?, and iii) Which individuals and organizations (authors, affiliated institutions, and funding bodies) have been involved in the development of AfricaNLP? We quantitatively examine the contributions of AfricaNLP research using 1.9K NLP paper abstracts, 4.9K author contributors, and 7.8K human-annotated contribution sentences (AfricaNLPContributions) along with benchmark results. Our dataset and continuously existing NLP progress tracking website provide a powerful lens for tracing AfricaNLP research trends and hold potential for generating data-driven literature surveys.
Authors: Yindong Wang, Martin Prei{\ss}, Margarita Bugue\~no, Jan Vincent Hoffbauer, Abdullatif Ghajar, Tolga Buz, Gerard de Melo
Abstract: Large Language Models (LLMs) frequently confabulate scientific facts, severely undermining their trustworthiness. Addressing this challenge requires benchmarks that go beyond binary factuality and enable fine-grained evaluation. We introduce ReFACT (Reddit False And Correct Texts), a benchmark of 1,001 expert-annotated question-answer pairs spanning diverse scientific domains for the detection of scientific confabulation. Each instance includes both a scientifically correct answer and a non-factual counterpart annotated with precise error spans and error types. ReFACT enables multi-stage evaluation: (1) confabulation detection, (2) fine-grained error localization, and (3) correction. We benchmark 9 state-of-the-art LLMs, revealing limited performance (about 50 percent accuracy). Even top models such as GPT-4o fail to distinguish factual from confabulated scientific answers, raising concerns about the reliability of LLM-as-judge evaluation paradigms. Our findings highlight the need for fine-grained, human-validated benchmarks to detect and correct scientific confabulation in domain-specific contexts. The dataset is available at: https://github.com/ddz5431/ReFACT
Authors: Xiaoyan Zhao, Ming Yan, Yang Zhang, Yang Deng, Jian Wang, Fengbin Zhu, Yilun Qiu, Hong Cheng, Tat-Seng Chua
Abstract: Conversational Recommender Systems (CRSs) aim to provide personalized recommendations through multi-turn natural language interactions with users. Given the strong interaction and reasoning skills of Large Language Models (LLMs), leveraging LLMs for CRSs has recently emerged as a promising direction. However, existing LLM-based methods often lack explicit optimization of interaction strategies, instead relying on unified prompts and the LLM's internal knowledge to decide how to interact, which can lead to suboptimal outcomes. In this paper, we propose a novel Reinforced Strategy Optimization (RSO) method for CRS, which decomposes the process of generating strategy-driven response decisions into the macro-level strategy planning and micro-level strategy adaptation through a network-of-experts architecture. At the macro level, a Planner expert selects macro-level interaction strategies (e.g., recommend, explain, encourage). At the micro level, an Actor expert generates detailed responses conditioned on the selected macro-level strategy, guided by auxiliary experts that provide complementary information such as user preferences and factual grounding. This hierarchical decomposition disentangles the optimization of different sub-tasks involved in CRS response generation, enabling more tractable learning at each level. To address the scarcity of high-quality multi-turn training data, we formulate strategy learning as a reinforcement learning problem, guided by an LLM-based reward model to achieve automatic strategy exploration. Extensive experiments show that RSO significantly improves interaction performance compared to state-of-the-art baselines, demonstrating the effectiveness of explicit hierarchical strategy optimization for CRS.
Authors: Mariia Fedorova, Andrey Kutuzov, Francesco Periti, Yves Scherrer
Abstract: We apply definition generators based on open-weights large language models to the task of creating explanations of novel senses, taking target word usages as an input. To this end, we employ the datasets from the AXOLOTL'24 shared task on explainable semantic change modeling, which features Finnish, Russian and German languages. We fine-tune and provide publicly the open-source models performing higher than the best submissions of the aforementioned shared task, which employed closed proprietary LLMs. In addition, we find that encoder-decoder definition generators perform on par with their decoder-only counterparts.
Authors: Jinyeop Song, Song Wang, Julian Shun, Yada Zhu
Abstract: Knowledge-graph retrieval-augmented generation (KG-RAG) couples large language models (LLMs) with structured, verifiable knowledge graphs (KGs) to reduce hallucinations and expose reasoning traces. However, many KG-RAG systems compose multiple LLM modules (e.g planning, reasoning, and responding), inflating inference cost and binding behavior to a specific target KG. To address this, we introduce KG-R1, an agentic KG retrieval-augmented generation (KG-RAG) framework through reinforcement learning (RL). KG-R1 utilizes a single agent that interacts with KGs as its environment, learning to retrieve at each step and incorporating the retrieved information into its reasoning and generation. The process is optimized through end-to-end RL. In controlled experiments across Knowledge-Graph Question Answering (KGQA) benchmarks, our method demonstrates both efficiency and transferability: Using Qwen-2.5-3B, KG-R1 improves answer accuracy with fewer generation tokens than prior multi-module workflow methods that use larger foundation or fine-tuned models. Furthermore, KG-R1 enables plug and play: after training, it maintains strong accuracy on new KGs without modification. These properties make KG-R1 a promising KG-RAG framework for real-world deployment. Our code is publicly available at https://github.com/Jinyeop3110/KG-R1.
Authors: Mainak Ghosh, Michael E. Rose, Sebastian Erhardt, Erik Buunk, Dietmar Harhoff
Abstract: PaECTER is an open-source document-level encoder specific for patents. We fine-tune BERT for Patents with examiner-added citation information to generate numerical representations for patent documents. PaECTER performs better in similarity tasks than current state-of-the-art models used in the patent domain. More specifically, our model outperforms the patent specific pre-trained language model (BERT for Patents) and general-purpose text embedding models (e.g., E5, GTE, and BGE) on our patent citation prediction test dataset on different rank evaluation metrics. PaECTER predicts at least one most similar patent at a rank of 1.32 on average when compared against 25 irrelevant patents. Numerical representations generated by PaECTER from patent text can be used for downstream tasks such as classification, tracing knowledge flows, or semantic similarity search. Semantic similarity search is especially relevant in the context of prior art search for both inventors and patent examiners.
Authors: Harsh Chaudhari, Giorgio Severi, John Abascal, Anshuman Suri, Matthew Jagielski, Christopher A. Choquette-Choo, Milad Nasr, Cristina Nita-Rotaru, Alina Oprea
Abstract: Retrieval Augmented Generation (RAG) expands the capabilities of modern large language models (LLMs), by anchoring, adapting, and personalizing their responses to the most relevant knowledge sources. It is particularly useful in chatbot applications, allowing developers to customize LLM output without expensive retraining. Despite their significant utility in various applications, RAG systems present new security risks. In this work, we propose a novel attack that allows an adversary to inject a single malicious document into a RAG system's knowledge base, and mount a backdoor poisoning attack. We design Phantom, a general two-stage optimization framework against RAG systems, that crafts a malicious poisoned document leading to an integrity violation in the model's output. First, the document is constructed to be retrieved only when a specific naturally occurring trigger sequence of tokens appears in the victim's queries. Second, the document is further optimized with crafted adversarial text that induces various adversarial objectives on the LLM output, including refusal to answer, reputation damage, privacy violations, and harmful behaviors.We demonstrate our attacks on multiple open-source LLM architectures, including Gemma, Vicuna, and Llama, and show that they transfer to closed-source models such as GPT-3.5 Turbo and GPT-4. Finally, we successfully demonstrate our attack on an end-to-end black-box production RAG system: NVIDIA's "Chat with RTX''.
Authors: Ruiping Ren, Xing Yao, Shu Cole, Haining Wang
Abstract: As large language models (LLMs) become increasingly integral to the hospitality and tourism industry, concerns about their fairness in serving diverse identity groups persist. Grounded in social identity theory and sociotechnical systems theory, this study examines ethnic and gender biases in travel recommendations generated by LLMs. Using fairness probing, we analyze outputs from three leading open-source LLMs. The results show that test accuracy for both ethnicity and gender classifiers exceed random chance. Analysis of the most influential features reveals the presence of stereotype bias in LLM-generated recommendations. We also found hallucinations among these features, occurring more frequently in recommendations for minority groups. These findings indicate that LLMs exhibit ethnic and gender bias when functioning as travel planning assistants. This study underscores the need for bias mitigation strategies to improve the inclusivity and reliability of generative AI-driven travel planning assistance.
Authors: Mohamed Ayoub Ben Ayad, Jelena Mitrovic, Michael Granitzer
Abstract: We introduce Krony-PT, a compression technique for GPT-2 based on Kronecker products. We specifically target the feed-forward weights of each transformer block, and systematically compress the feed-forward layer matrices to various degrees. We introduce a modified Van Loan decomposition to initialize new Kronecker factors, and also propose a new pruning-based initialization technique. Our method compresses the original 124M-parameter GPT-2 to various smaller models, ranging from 80M to 96M. Our 81M model variant outperforms DistilGPT2 on next-token prediction across all standard language modeling datasets, and shows competitive or comparable performance with significantly larger Kronecker-based compressions of GPT-2.
Authors: Kaiyuan Hou, Minghui Zhao, Lilin Xu, Yuang Fan, Xiaofan Jiang
Abstract: Top-down images play an important role in safety-critical settings such as autonomous navigation and aerial surveillance, where they provide holistic spatial information that front-view images cannot capture. Despite this, Vision Language Models (VLMs) are mostly trained and evaluated on front-view benchmarks, leaving their performance in the top-down setting poorly understood. Existing evaluations also overlook a unique property of top-down images: their physical meaning is preserved under rotation. In addition, conventional accuracy metrics can be misleading, since they are often inflated by hallucinations or "lucky guesses", which obscures a model's true reliability and its grounding in visual evidence. To address these issues, we introduce TDBench, a benchmark for top-down image understanding that includes 2000 curated questions for each rotation. We further propose RotationalEval (RE), which measures whether models provide consistent answers across four rotated views of the same scene, and we develop a reliability framework that separates genuine knowledge from chance. Finally, we conduct four case studies targeting underexplored real-world challenges. By combining rigorous evaluation with reliability metrics, TDBench not only benchmarks VLMs in top-down perception but also provides a new perspective on trustworthiness, guiding the development of more robust and grounded AI systems. Project homepage: https://github.com/Columbia-ICSL/TDBench
Authors: Run Luo, Lu Wang, Wanwei He, Longze Chen, Jiaming Li, Xiaobo Xia
Abstract: Existing efforts in building Graphical User Interface (GUI) agents largely rely on the training paradigm of supervised fine-tuning on Large Vision-Language Models (LVLMs). However, this approach not only demands extensive amounts of training data but also struggles to effectively understand GUI screenshots and generalize to unseen interfaces. The issue significantly limits its application in real-world scenarios, especially for high-level tasks. Inspired by Reinforcement Fine-Tuning (RFT) in large reasoning models (e.g., DeepSeek-R1), which efficiently enhances the problem-solving capabilities of large language models in real-world settings, we propose \name, the first reinforcement learning framework designed to enhance the GUI capabilities of LVLMs in high-level real-world task scenarios, through unified action space rule modeling. By leveraging a small amount of carefully curated high-quality data across multiple platforms (including Windows, Linux, MacOS, Android, and Web) and employing policy optimization algorithms such as Group Relative Policy Optimization (GRPO) to update the model, \name achieves superior performance using only 0.02\% of the data (3K vs. 13M) compared to previous state-of-the-art methods like OS-Atlas across eight benchmarks spanning three different platforms (mobile, desktop, and web). These results demonstrate the immense potential of reinforcement learning based on unified action space rule modeling in improving the execution capabilities of LVLMs for real-world GUI agent tasks.
Authors: Yixuan Even Xu, Yash Savani, Fei Fang, J. Zico Kolter
Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as the leading approach for enhancing reasoning capabilities in large language models. However, it faces a fundamental compute and memory asymmetry: rollout generation is embarrassingly parallel and memory-light, whereas policy updates are communication-heavy and memory-intensive. To address this, we introduce PODS (Policy Optimization with Down-Sampling), which decouples rollout generation from policy updates by training only on a strategically selected subset of rollouts, maintaining learning quality while dramatically reducing update costs. We propose a principled subset selection criterion, max-variance down-sampling, that maximizes reward diversity, and provide an efficient $O(n\log n)$ implementation. Empirically, Group Relative Policy Optimization (GRPO) with PODS achieves the peak test accuracy of vanilla GRPO at least $\mathbf{1.7\times}$ faster across the different reasoning benchmarks and hardware configurations we tested.
Authors: Peter Chen, Xiaopeng Li, Ziniu Li, Xi Chen, Tianyi Lin
Abstract: Reinforcement learning (RL) has proven effective in strengthening the reasoning capabilities of large language models (LLMs). A widely adopted method, Group Relative Policy Optimization (GRPO), has shown strong empirical results in training DeepSeek-R1. However, GRPO fails to update the policy when all responses within a group are incorrect (i.e., \emph{all-negative-sample} groups). This limitation underscores a key gap between artificial and human intelligence: unlike humans, who can learn from mistakes, GRPO discards these signals. Our first contribution is to introduce a simple framework that mitigates the all-negative-sample issue by incorporating response diversity within groups using a \textit{step-wise} judge model, which can be either directly trained or adapted from existing LLMs. We prove that this diversification can accelerate GRPO's learning dynamics in a simplified setting. We also empirically validate the proposed stepwise guided policy optimization (SGPO) method, demonstrating consistent gains across model sizes (7B, 14B, 32B) in offline and online training on 9 benchmarks, including base and distilled variants. Our results highlight two advantages: (i) SGPO surpasses GRPO, especially in the early and mid-training stages where all-negative-sample groups are prevalent; and (ii) SGPO does not require judge models to generate correct answers, differentiating it from knowledge distillation methods.
Authors: Sen Fang, Weiyuan Ding, Bowen Xu
Abstract: Evaluating the programming robustness of large language models (LLMs) is paramount for ensuring their reliability in AI-based software development. However, adversarial attacks exhibit fundamental limitations that compromise fair robustness assessment: they demonstrate contradictory evaluation outcomes where different attack strategies tend to favor different models, and more critically, they operate solely through external perturbations, failing to capture the intrinsic stability essential for autonomous coding agents where subsequent inputs are endogenously generated by the model itself. We introduce EVALOOOP, a novel assessment framework that evaluates robustness from a self-consistency perspective, leveraging the natural duality inherent in software engineering tasks (e.g., code generation and code summarization). EVALOOOP establishes a self-contained feedback loop where an LLM iteratively transforms between code and natural language until functional failure occurs, with robustness quantified by a novel Average Sustainable Loops (ASL) metric-the mean number of iterations maintaining functional correctness across benchmark tasks. This cyclical strategy intrinsically evaluates robustness without relying on external attack configurations, providing a unified metric that reveals how effectively LLMs preserve semantic integrity through sustained self-referential transformations. We evaluate 96 popular LLMs, ranging from 0.5B to 685B parameters, on EVALOOOP equipped with the MBPP Plus benchmark, and found that EVALOOOP typically induces a 2.65%-47.62% absolute drop in pass@1 accuracy within ten loops. Intriguingly, robustness does not always align with initial performance (i.e., one-time query); for instance, Qwen3-235B-A22B-Instruct-2507, despite inferior initial code generation compared to OpenAI's o-series models and DeepSeek-V3, demonstrated the superior robustness (ASL score).
Authors: Eric Hanchen Jiang, Haozheng Luo, Shengyuan Pang, Xiaomin Li, Zhenting Qi, Hengli Li, Cheng-Fu Yang, Zongyu Lin, Xinfeng Li, Hao Xu, Kai-Wei Chang, Ying Nian Wu
Abstract: Large Language Models (LLMs) struggle with reliable mathematical reasoning, and current verification methods are often computationally expensive. This paper introduces the Energy Outcome Reward Model (EORM), a highly efficient, lightweight post-hoc verifier designed to address this challenge. EORM uses an energy-based framework to rank Chain-of-Thought (CoT) solutions, learning to distinguish correct from incorrect reasoning using only simple outcome labels, thus eliminating the need for expensive annotations. With only 55M parameters, over 127 times smaller than typical reward models, EORM boosts the accuracy of Llama 3 8B to 90.7\% on GSM8k and 63.7\% on MATH. This performance is achieved by efficiently selecting the optimal reasoning path from a pool of candidates, allowing it to match or exceed the accuracy of far more resource-intensive Best-of-N sampling techniques. Crucially, our experiments show that EORM generalizes effectively to out-of-distribution problems and unseen models, indicating it learns fundamental principles of valid reasoning. This robustness, combined with its efficiency, establishes EORM as a practical tool for deploying more dependable LLMs in complex, real-world applications.
Authors: Chih-Kai Yang, Neo S. Ho, Hung-yi Lee
Abstract: With advancements in large audio-language models (LALMs), which enhance large language models (LLMs) with auditory capabilities, these models are expected to demonstrate universal proficiency across various auditory tasks. While numerous benchmarks have emerged to assess LALMs' performance, they remain fragmented and lack a structured taxonomy. To bridge this gap, we conduct a comprehensive survey and propose a systematic taxonomy for LALM evaluations, categorizing them into four dimensions based on their objectives: (1) General Auditory Awareness and Processing, (2) Knowledge and Reasoning, (3) Dialogue-oriented Ability, and (4) Fairness, Safety, and Trustworthiness. We provide detailed overviews within each category and highlight challenges in this field, offering insights into promising future directions. To the best of our knowledge, this is the first survey specifically focused on the evaluations of LALMs, providing clear guidelines for the community. We will release the collection of the surveyed papers and actively maintain it to support ongoing advancements in the field.
Authors: Kai Sun, Yushi Bai, Zhen Yang, Jiajie Zhang, Ji Qi, Lei Hou, Juanzi Li
Abstract: Large Multimodal Models (LMMs) typically build on ViTs (e.g., CLIP), yet their training with simple random in-batch negatives limits the ability to capture fine-grained visual differences, particularly in geometric scenarios. To address this challenge, we propose a novel hard negative contrastive learning framework for the vision encoder, which combines image-based contrastive learning using generation-based hard negatives created by perturbing diagram generation code, and text-based contrastive learning using rule-based negatives derived from modified geometric descriptions and retrieval-based negatives selected based on caption similarity. We train a vision encoder (CLIP) using our hard negative training method, namely MMCLIP (Multimodal Math CLIP), and subsequently train an LMM for geometric problem-solving. Experiments show that our trained model, MMGeoLM, significantly outperforms other open-source models on three geometric reasoning benchmarks. Even with a size of 7B, it can rival powerful closed-source models like GPT-4o. We further conduct ablation studies to analyze three key factors: hard negative types, the efficiency of image-based negatives, and training configurations. These analyses yield important insights into optimizing the training pipeline of vision encoder for fine-grained geometric reasoning tasks. https://github.com/THU-KEG/MMGeoLM.
Authors: Akshat Naik, Patrick Quinn, Guillermo Bosch, Emma Goun\'e, Francisco Javier Campos Zabala, Jason Ross Brown, Edward James Young
Abstract: As Large Language Model (LLM) agents become more widespread, associated misalignment risks increase. While prior research has studied agents' ability to produce harmful outputs or follow malicious instructions, it remains unclear how likely agents are to spontaneously pursue unintended goals in realistic deployments. In this work, we approach misalignment as a conflict between the internal goals pursued by the model and the goals intended by its deployer. We introduce a misalignment propensity benchmark, \textsc{AgentMisalignment}, a benchmark suite designed to evaluate the propensity of LLM agents to misalign in realistic scenarios. Evaluations cover behaviours such as avoiding oversight, resisting shutdown, sandbagging, and power-seeking. Testing frontier models, we find that more capable agents tend to exhibit higher misalignment on average. We also systematically vary agent personalities through different system prompts and observe that persona characteristics can strongly and unpredictably influence misalignment, sometimes more than the choice of model itself. Our results reveal the limitations of current alignment methods for autonomous LLM agents and underscore the need to rethink misalignment in realistic deployment settings.
Authors: Marcel Mateos Salles, Praney Goyal, Pradyut Sekhsaria, Hai Huang, Randall Balestriero
Abstract: Large Language Models (LLMs) are commonly finetuned for a variety of use cases and domains. A common approach is to leverage Low-Rank Adaptation (LoRA) -- known to provide strong performance at low resource costs. In this study, we demonstrate that LoRA actually opens the door to short-cut vulnerabilities -- and the more resource efficient is the LoRA setup, the more vulnerable will be the finetuned model to aggressive attacks. To measure that vulnerability, we introduce Seamless Spurious Token Injection (SSTI), where we find that LoRA exclusively focuses on even just a single token that is spuriously correlated with downstream labels. In short, injection of that spurious token during finetuning ensure that the model's prediction at test-time can be manipulated on-demand. We conducted experiments across model families and datasets to evaluate the impact of SSTI during LoRA finetuning while providing possible mitigations. Our experiments conclude that none of the existing checkers and preprocessors can sanitize a dataset raising new concerns for data quality and AI safety.
Authors: Siao Tang, Xinyin Ma, Gongfan Fang, Xinchao Wang
Abstract: Recent advancements in large reasoning models (LRMs) like DeepSeek-R1 and OpenAI o1 series have achieved notable performance enhancements on complex reasoning tasks by scaling up the generation length by Chain-of-Thought (CoT). However, a critical issue is their tendency to produce excessively verbose reasoning processes, leading to the inefficiency problem. Existing literature on improving efficiency mainly adheres to the before-reasoning paradigms such as prompting and reasoning or fine-tuning and reasoning, but ignores the promising direction of directly encouraging the model to speak concisely by intervening during the generation of reasoning. In order to fill the blank, we propose a framework dubbed ConciseHint, which continuously encourages the reasoning model to speak concisely by injecting learnable hints (manually designed or learned on concise data) during the generation of the reasoning. Besides, ConciseHint is adaptive to the complexity of the query by adaptively adjusting the hint intensity, which ensures it will not undermine model performance. Experiments on the state-of-the-art LRMs, including DeepSeek-R1 and Qwen-3 series, demonstrate that our method can effectively produce concise reasoning while maintaining the performance well. Moreover, we show that ConciseHint is flexible and can be seamlessly integrated with existing methods to further push the upper bound of the efficiency.
Authors: Vladimir Bogachev, Vladimir Aletov, Alexander Molozhavenko, Denis Bobkov, Vera Soboleva, Aibek Alanov, Maxim Rakhuba
Abstract: This work presents a novel, fully Riemannian framework for Low-Rank Adaptation (LoRA) that geometrically treats low-rank adapters by optimizing them directly on the fixed-rank manifold. This formulation eliminates the parametrization ambiguity present in standard Euclidean optimizers. Our framework integrates three key components to achieve this: (1) we derive Riemannion, a new Riemannian optimizer on the fixed-rank matrix manifold that generalizes the recently proposed Muon optimizer; (2) we develop a Riemannian gradient-informed LoRA initialization, and (3) we provide an efficient implementation without prominent overhead that uses automatic differentiation to compute arising geometric operations while adhering to best practices in numerical linear algebra. Comprehensive experimental results on both LLM and diffusion model architectures demonstrate that our approach yields consistent and noticeable improvements in convergence speed and final task performance over both standard LoRA and its state-of-the-art modifications.
Authors: Yuqi Pan, Yongqi An, Zheng Li, Yuhong Chou, Ruijie Zhu, Xiaohui Wang, Mingxuan Wang, Jinqiao Wang, Guoqi Li
Abstract: The Transformer architecture, despite its widespread success, struggles with long-context scenarios due to quadratic computation and linear memory growth. While various linear attention variants mitigate these efficiency constraints by compressing context into fixed-size states, they often degrade performance in tasks such as in-context retrieval and reasoning. To address this limitation and achieve more effective context compression, we propose two key innovations. First, we introduce a row-sparse update formulation for linear attention by conceptualizing state updating as information classification. This enables sparse state updates via softmax-based top-$k$ hard classification, thereby extending receptive fields and reducing inter-class interference. Second, we present Sparse State Expansion (SSE) within the sparse framework, which expands the contextual state into multiple partitions, effectively decoupling parameter size from state capacity while maintaining the sparse classification paradigm. Supported by efficient parallelized implementations, our design achieves effective classification and highly discriminative state representations. We extensively validate SSE in both pure linear and hybrid (SSE-H) architectures across language modeling, in-context retrieval, and mathematical reasoning benchmarks. SSE demonstrates strong retrieval performance and scales favorably with state size. Moreover, after reinforcement learning (RL) training, our 2B SSE-H model achieves state-of-the-art mathematical reasoning performance among small reasoning models, scoring 64.5 on AIME24 and 50.2 on AIME25, significantly outperforming similarly sized open-source Transformers. These results highlight SSE as a promising and efficient architecture for long-context modeling.
Authors: Tianyi Ma, Yue Zhang, Zehao Wang, Parisa Kordjamshidi
Abstract: Vision-and-Language Navigation (VLN) poses significant challenges for agents to interpret natural language instructions and navigate complex 3D environments. While recent progress has been driven by large-scale pre-training and data augmentation, current methods still struggle to generalize to unseen scenarios, particularly when complex spatial and temporal reasoning is required. In this work, we propose SkillNav, a modular framework that introduces structured, skill-based reasoning into Transformer-based VLN agents. Our method decomposes navigation into a set of interpretable atomic skills (e.g., Vertical Movement, Area and Region Identification, Stop and Pause), each handled by a specialized agent. To support targeted skill training without manual data annotation, we construct a synthetic dataset pipeline that generates diverse, linguistically natural, skill-specific instruction-trajectory pairs. We then introduce a novel training-free Vision-Language Model (VLM)-based router, which dynamically selects the most suitable agent at each time step by aligning sub-goals with visual observations and historical actions. SkillNav obtains competitive results on commonly used benchmarks and establishes state-of-the-art generalization to the GSA-R2R, a benchmark with novel instruction styles and unseen environments.
Authors: Shan Wang, Maying Shen, Nadine Chang, Chuong Nguyen, Hongdong Li, Jose M. Alvarez
Abstract: Multimodal large language models achieve strong performance across diverse tasks but remain prone to hallucinations, where outputs are not grounded in visual inputs. This issue can be attributed to two main biases: text-visual bias, the overreliance on prompts and prior outputs, and co-occurrence bias, spurious correlations between frequently paired objects. We propose Gradient-based Influence-Aware Constrained Decoding (GACD), an inference-based method, that addresses both biases without auxiliary models, and is readily applicable to existing models without finetuning. The core of our approach is bias estimation, which uses first-order Taylor gradients to understand the contribution of individual tokens-visual features and text tokens-to the current output. Based on this analysis, GACD mitigates hallucinations through two components: (1) suppressing spurious visual features correlated with the output objects, and (2) rebalancing cross-modal contributions by strengthening visual features relative to text. Experiments across multiple benchmarks demonstrate that GACD effectively reduces hallucinations and improves the visual grounding of MLLM outputs.
Authors: Mohamed Mohamed, Brennan Nichyporuk, Douglas L. Arnold, Tal Arbel
Abstract: Vision-language models have demonstrated impressive capabilities in generating 2D images under various conditions; however, the success of these models is largely enabled by extensive, readily available pretrained foundation models. Critically, comparable pretrained models do not exist for 3D, significantly limiting progress. As a result, the potential of vision-language models to produce high-resolution 3D counterfactual medical images conditioned solely on natural language remains unexplored. Addressing this gap would enable powerful clinical and research applications, such as personalized counterfactual explanations, simulation of disease progression, and enhanced medical training by visualizing hypothetical conditions in realistic detail. Our work takes a step toward this challenge by introducing a framework capable of generating high-resolution 3D counterfactual medical images of synthesized patients guided by free-form language prompts. We adapt state-of-the-art 3D diffusion models with enhancements from Simple Diffusion and incorporate augmented conditioning to improve text alignment and image quality. To our knowledge, this is the first demonstration of a language-guided native-3D diffusion model applied to neurological imaging, where faithful three-dimensional modeling is essential. On two neurological MRI datasets, our framework simulates varying counterfactual lesion loads in Multiple Sclerosis and cognitive states in Alzheimer's disease, generating high-quality images while preserving subject fidelity. Our results lay the groundwork for prompt-driven disease progression analysis in 3D medical imaging. Project link - https://lesupermomo.github.io/imagining-alternatives/.
URLs: https://lesupermomo.github.io/imagining-alternatives/.
Authors: Haywood Gelman, John D. Hastings, David Kenley
Abstract: Insider threats are a growing organizational problem due to the complexity of identifying their technical and behavioral elements. A large research body is dedicated to the study of insider threats from technological, psychological, and educational perspectives. However, research in this domain has been generally dependent on datasets that are static and limited access which restricts the development of adaptive detection models. This study introduces a novel, ethically grounded approach that uses the large language model (LLM) Claude Sonnet 3.7 to dynamically synthesize syslog messages, some of which contain indicators of insider threat scenarios. The messages reflect real-world data distributions by being highly imbalanced (1% insider threats). The syslogs were analyzed for insider threats by both Sonnet 3.7 and GPT-4o, with their performance evaluated through statistical metrics including accuracy, precision, recall, F1, specificity, FAR, MCC, and ROC AUC. Sonnet 3.7 consistently outperformed GPT-4o across nearly all metrics, particularly in reducing false alarms and improving detection accuracy. The results show strong promise for the use of LLMs in synthetic dataset generation and insider threat detection.
Authors: Yifan Zhang
Abstract: We give a simple and provably correct replacement for the contested ``domain-extension'' in Step 9 of a recent windowed-QFT lattice algorithm with complex-Gaussian windows~\citep{chen2024quantum}. As acknowledged by the author, the reported issue is due to a periodicity/support mismatch when applying domain extension to only the first coordinate in the presence of offsets. Our drop-in subroutine replaces domain extension by a pair-shift difference that cancels all unknown offsets exactly and synthesizes a uniform cyclic subgroup (a zero-offset coset) of order $P$ inside $(\mathbb{Z}_{M_2})^n$. A subsequent QFT enforces the intended modular linear relation by plain character orthogonality. The sole structural assumption is a residue-accessibility condition enabling coherent auxiliary cleanup; no amplitude periodicity is used. The unitary is reversible, uses $\mathrm{poly}(\log M_2)$ gates, and preserves upstream asymptotics.
Authors: Yujun Zhou, Zhenwen Liang, Haolin Liu, Wenhao Yu, Kishan Panaganti, Linfeng Song, Dian Yu, Xiangliang Zhang, Haitao Mi, Dong Yu
Abstract: Large language models (LLMs) are increasingly trained with reinforcement learning from verifiable rewards (RLVR), yet real-world deployment demands models that can self-improve without labels or external judges. Existing self-improvement approaches primarily rely on self-confirmation signals (e.g., confidence, entropy, or consistency) to generate rewards. This reliance drives models toward over-confident, majority-favored solutions, causing an entropy collapse that degrades pass@n and reasoning complexity. To address this, we propose EVOL-RL, a label-free framework that mirrors the evolutionary principle of balancing selection with variation. Concretely, EVOL-RL retains the majority-voted answer as an anchor for stability, but adds a novelty-aware reward that scores each sampled solution by how different its reasoning is from other concurrently generated responses. This majority-for-stability + novelty-for-exploration rule mirrors the variation-selection principle: selection prevents drift, while novelty prevents collapse. Evaluation results show that EVOL-RL consistently outperforms the majority-only baseline; e.g., training on label-free AIME24 lifts Qwen3-4B-Base AIME25 pass@1 from baseline's 4.6% to 16.4%, and pass@16 from 18.5% to 37.9%. EVOL-RL not only prevents in-domain diversity collapse but also improves out-of-domain generalization (from math reasoning to broader tasks, e.g., GPQA, MMLU-Pro, and BBEH). The code is available at: https://github.com/YujunZhou/EVOL-RL.
Authors: Yu Gu, Jingjing Fu, Xiaodong Liu, Jeya Maria Jose Valanarasu, Noel CF Codella, Reuben Tan, Qianchu Liu, Ying Jin, Sheng Zhang, Jinyu Wang, Rui Wang, Lei Song, Guanghui Qin, Naoto Usuyama, Cliff Wong, Hao Cheng, Hohin Lee, Praneeth Sanapathi, Sarah Hilado, Jiang Bian, Javier Alvarez-Valle, Mu Wei, Khalil Malik, Jianfeng Gao, Eric Horvitz, Matthew P Lungren, Hoifung Poon, Paul Vozila
Abstract: Large frontier models like GPT-5 now achieve top scores on medical benchmarks. But our stress tests tell a different story. Leading systems often guess correctly even when key inputs like images are removed, flip answers under trivial prompt changes, and fabricate convincing yet flawed reasoning. These aren't glitches; they expose how today's benchmarks reward test-taking tricks over medical understanding. We evaluate six flagship models across six widely used benchmarks and find that high leaderboard scores hide brittleness and shortcut learning. Through clinician-guided rubric evaluation, we show that benchmarks vary widely in what they truly measure yet are treated interchangeably, masking failure modes. We caution that medical benchmark scores do not directly reflect real-world readiness. If we want AI to earn trust in healthcare, we must demand more than leaderboard wins and must hold systems accountable for robustness, sound reasoning, and alignment with real medical demands.
Authors: Jiakai Tang, Yujie Luo, Xunke Xi, Fei Sun, Xueyang Feng, Sunhao Dai, Chao Yi, Dian Chen, Zhujin Gao, Yang Li, Xu Chen, Wen Chen, Jian Wu, Yuning Jiang, Bo Zheng
Abstract: Traditional recommender systems rely on passive feedback mechanisms that limit users to simple choices such as like and dislike. However, these coarse-grained signals fail to capture users' nuanced behavior motivations and intentions. In turn, current systems cannot also distinguish which specific item attributes drive user satisfaction or dissatisfaction, resulting in inaccurate preference modeling. These fundamental limitations create a persistent gap between user intentions and system interpretations, ultimately undermining user satisfaction and harming system effectiveness. To address these limitations, we introduce the Interactive Recommendation Feed (IRF), a pioneering paradigm that enables natural language commands within mainstream recommendation feeds. Unlike traditional systems that confine users to passive implicit behavioral influence, IRF empowers active explicit control over recommendation policies through real-time linguistic commands. To support this paradigm, we develop RecBot, a dual-agent architecture where a Parser Agent transforms linguistic expressions into structured preferences and a Planner Agent dynamically orchestrates adaptive tool chains for on-the-fly policy adjustment. To enable practical deployment, we employ simulation-augmented knowledge distillation to achieve efficient performance while maintaining strong reasoning capabilities. Through extensive offline and long-term online experiments, RecBot shows significant improvements in both user satisfaction and business outcomes.
Authors: Xingyu Fu, Siyi Liu, Yinuo Xu, Pan Lu, Guangqiuse Hu, Tianbo Yang, Taran Anantasagar, Christopher Shen, Yikai Mao, Yuanzhe Liu, Keyush Shah, Chung Un Lee, Yejin Choi, James Zou, Dan Roth, Chris Callison-Burch
Abstract: Can humans identify AI-generated (fake) videos and provide grounded reasons? While video generation models have advanced rapidly, a critical dimension -- whether humans can detect deepfake traces within a generated video, i.e., spatiotemporal grounded visual artifacts that reveal a video as machine generated -- has been largely overlooked. We introduce DeeptraceReward, the first fine-grained, spatially- and temporally- aware benchmark that annotates human-perceived fake traces for video generation reward. The dataset comprises 4.3K detailed annotations across 3.3K high-quality generated videos. Each annotation provides a natural-language explanation, pinpoints a bounding-box region containing the perceived trace, and marks precise onset and offset timestamps. We consolidate these annotations into 9 major categories of deepfake traces that lead humans to identify a video as AI-generated, and train multimodal language models (LMs) as reward models to mimic human judgments and localizations. On DeeptraceReward, our 7B reward model outperforms GPT-5 by 34.7% on average across fake clue identification, grounding, and explanation. Interestingly, we observe a consistent difficulty gradient: binary fake v.s. real classification is substantially easier than fine-grained deepfake trace detection; within the latter, performance degrades from natural language explanations (easiest), to spatial grounding, to temporal labeling (hardest). By foregrounding human-perceived deepfake traces, DeeptraceReward provides a rigorous testbed and training signal for socially aware and trustworthy video generation.
Authors: Runyan Tan, Shuang Wu, Phillip Howard
Abstract: Obtaining high-quality outputs from Large Language Models (LLMs) often depends upon the choice of a sampling-based decoding strategy to probabilistically choose the next token at each generation step. While a variety of such sampling methods have been proposed, their performance can be sensitive to the selection of hyperparameters which may require different settings depending upon the generation task and temperature configuration. In this work, we introduce $p$-less sampling: an information-theoretic approach to sampling which dynamically sets a truncation threshold at each decoding step based on the entire token probability distribution. Unlike existing methods, $p$-less sampling has no hyperparameters and consistently produces high-quality outputs as temperature increases. We provide theoretical perspectives on $p$-less sampling to ground our proposed method and conduct experiments to empirically validate its effectiveness across a range of math, logical reasoning, and creative writing tasks. Our results demonstrate how $p$-less sampling consistently outperforms existing sampling approaches while exhibiting much less degradation in text quality at higher temperature values. We further show how $p$-less achieves greater inference-time efficiency than alternative methods through lower average token sampling times and shorter generation lengths, without sacrificing accuracy. Finally, we provide analyses to highlight the benefits of $p$-less through qualitative examples, case studies, and diversity assessments.
Authors: Siyang Wu, Honglin Bao, Sida Li, Ari Holtzman, James A. Evans
Abstract: We develop signatures of capacity familiarity to characterize large language model (LLM) benchmarks and their meaningful overlaps. Benchmark signatures probe the capacity required for benchmark performance. We formally define them as a set of salient tokens drawn from in-the-wild, naturally authored corpora, where LLM token perplexity, reflecting more or less pre-training exposure, becomes highly predictive of LLM benchmark performance. Through a large-scale meta-evaluation, we extract benchmark signatures via stepwise forward selection with linear regressions across 32 LLMs and 88 benchmarks spanning diverse knowledge, coding, logic, instruction following, math, language, reasoning, and world modeling. Our analysis situates signatures in relation to both the semantic similarity of benchmark questions and the correlation of model performance. While performance overlaps are universally high and semantic overlaps remain confined to a narrow mid-range, benchmark signatures prove highly informative in capturing variation, overlap, and divergence. We observe overlap in knowledge and reasoning subtasks, whereas multilingual and cultural benchmarks exhibit less similarity, even compared to cross-task overlap. Notably, performance-level results are strongly influenced by benchmark-orthogonal factors such as question format, highlighting limitations in LLM generalization, the conflation of performance with ability, and issues inherent in current mainstream benchmark agreement studies. Benchmark signatures, however, remain robust to such effects. Ultimately, we identify cross-functional overlaps across logic, math, language, instruction following, and world modeling, with coding emerging as the least overlapping domain. Together, these findings provide mechanistic insights into benchmark validity and LLM sensitivities, and sketch the underlying landscape of interconnected LLM capabilities.
Authors: Jianshuo Dong, Sheng Guo, Hao Wang, Zhuotao Liu, Tianwei Zhang, Ke Xu, Minlie Huang, Han Qiu
Abstract: Search agents connect LLMs to the Internet, enabling access to broader and more up-to-date information. However, unreliable search results may also pose safety threats to end users, establishing a new threat surface. In this work, we conduct two in-the-wild experiments to demonstrate both the prevalence of low-quality search results and their potential to misguide agent behaviors. To counter this threat, we introduce an automated red-teaming framework that is systematic, scalable, and cost-efficient, enabling lightweight and harmless safety assessments of search agents. Building on this framework, we construct the SafeSearch benchmark, which includes 300 test cases covering five categories of risks (e.g., misinformation and indirect prompt injection). Using this benchmark, we evaluate three representative search agent scaffolds, covering search workflow, tool-calling, and deep research, across 7 proprietary and 8 open-source backend LLMs. Our results reveal substantial vulnerabilities of LLM-based search agents: when exposed to unreliable websites, the highest ASR reached 90.5% for GPT-4.1-mini under a search workflow setting. Moreover, our analysis highlights the limited effectiveness of common defense practices, such as reminder prompting. This emphasizes the value of our framework in promoting transparency for safer agent development. Our codebase and test cases are publicly available: https://github.com/jianshuod/SafeSearch.
Authors: Fanding Huang, Guanbo Huang, Xiao Fan, Yi He, Xiao Liang, Xiao Chen, Qinting Jiang, Faisal Nadeem Khan, Jingyan Jiang, Zhi Wang
Abstract: A prevailing view in Reinforcement Learning for Verifiable Rewards (RLVR) interprets recent progress through the lens of an exploration-exploitation trade-off, a perspective largely shaped by token-level metrics. We re-examine this perspective, proposing that this perceived trade-off may not be a fundamental constraint but rather an artifact of the measurement level. To investigate this, we shift the analysis to the semantically rich hidden-state space, adopting Effective Rank (ER) to quantify exploration and proposing its novel first- and second-order derivatives, named Effective Rank Velocity (ERV) and Effective Rank Acceleration (ERA), to capture exploitation dynamics. Our analysis reveals that at the hidden-state level, exploration and exploitation could be decoupled (Sec. 4). This finding reveals an opportunity to enhance both capacities simultaneously. This insight motivates our method, Velocity-Exploiting Rank-Learning (VERL), the first to operationalize the principle of synergistic exploration-exploitation enhancement by directly shaping the RL advantage function. The key innovation is leveraging the theoretically stable ERA as a predictive meta-controller to create a synergistic, dual-channel incentive structure. Instead of forcing a trade-off, VERL prospectively amplifies rewards for exploration to preempt overconfidence and reinforces exploitative gains to consolidate reasoning. Experiments across diverse LLMs and reasoning benchmarks show consistent gains, including up to 21.4% absolute accuracy improvement on the challenging Gaokao 2024 dataset.
Authors: Yuepeng Hu, Zhengyuan Jiang, Mengyuan Li, Osama Ahmed, Zhicong Huang, Cheng Hong, Neil Gong
Abstract: Large language models (LLMs) are often modified after release through post-processing such as post-training or quantization, which makes it challenging to determine whether one model is derived from another. Existing provenance detection methods have two main limitations: (1) they embed signals into the base model before release, which is infeasible for already published models, or (2) they compare outputs across models using hand-crafted or random prompts, which are not robust to post-processing. In this work, we propose LLMPrint, a novel detection framework that constructs fingerprints by exploiting LLMs' inherent vulnerability to prompt injection. Our key insight is that by optimizing fingerprint prompts to enforce consistent token preferences, we can obtain fingerprints that are both unique to the base model and robust to post-processing. We further develop a unified verification procedure that applies to both gray-box and black-box settings, with statistical guarantees. We evaluate LLMPrint on five base models and around 700 post-trained or quantized variants. Our results show that LLMPrint achieves high true positive rates while keeping false positive rates near zero.
Authors: Fang Wu, Weihao Xuan, Heli Qi, Ximing Lu, Aaron Tu, Li Erran Li, Yejin Choi
Abstract: Although RLVR has become an essential component for developing advanced reasoning skills in LLMs, contemporary studies have documented training plateaus that emerge following thousands of optimization steps, demonstrating notable decreases in performance gains despite increased computational investment. This limitation stems from the sparse exploration patterns inherent in current RLVR practices, where models rely on limited rollouts that often miss critical reasoning paths and fail to provide systematic coverage of the solution space. We present DeepSearch, a framework that integrates Monte Carlo Tree Search directly into RLVR training. In contrast to existing methods that rely on tree search only at inference, DeepSearch embeds structured search into the training loop, enabling systematic exploration and fine-grained credit assignment across reasoning steps. Through training-time exploration, DeepSearch addresses the fundamental bottleneck of insufficient exploration, which leads to diminishing performance improvements over prolonged training steps. Our contributions include: (1) a global frontier selection strategy that prioritizes promising nodes across the search tree, (2) selection with entropy-based guidance that identifies confident paths for supervision, and (3) adaptive replay buffer training with solution caching for efficiency. Experiments on mathematical reasoning benchmarks show that DeepSearch achieves 62.95% average accuracy and establishes a new state-of-the-art for 1.5B reasoning models - using 5.7x fewer GPU hours than extended training approaches. These results highlight the importance of strategic exploration over brute-force scaling and demonstrate the promise of algorithmic innovation for advancing RLVR methodologies. DeepSearch establishes a new direction for scaling reasoning capabilities through systematic search rather than prolonged computation.
Authors: Arvind Murari Vepa, Yannan Yu, Jingru Gan, Anthony Cuturrufo, Weikai Li, Wei Wang, Fabien Scalzo, Yizhou Sun
Abstract: We introduce mpLLM, a prompt-conditioned hierarchical mixture-of-experts (MoE) architecture for visual question answering over multi-parametric 3D brain MRI (mpMRI). mpLLM routes across modality-level and token-level projection experts to fuse multiple interrelated 3D modalities, enabling efficient training without image-report pretraining. To address limited image-text paired supervision, mpLLM integrates a synthetic visual question answering (VQA) protocol that generates medically relevant VQA from segmentation annotations, and we collaborate with medical experts for clinical validation. mpLLM outperforms strong medical VLM baselines by 5.3% on average across multiple mpMRI datasets. Our study features three main contributions: (1) the first clinically validated VQA dataset for 3D brain mpMRI, (2) a novel multimodal LLM that handles multiple interrelated 3D modalities, and (3) strong empirical results that demonstrate the medical utility of our methodology. Ablations highlight the importance of modality-level and token-level experts and prompt-conditioned routing.
Authors: William Walden, Marc Mason, Orion Weller, Laura Dietz, Hannah Recknor, Bryan Li, Gabrielle Kaili-May Liu, Yu Hou, James Mayfield, Eugene Yang
Abstract: Generation of long-form, citation-backed reports is a primary use case for retrieval augmented generation (RAG) systems. While open-source evaluation tools exist for various RAG tasks, ones tailored to report generation are lacking. Accordingly, we introduce Auto-ARGUE, a robust LLM-based implementation of the recent ARGUE framework for report generation evaluation. We present analysis of Auto-ARGUE on the report generation pilot task from the TREC 2024 NeuCLIR track, showing good system-level correlations with human judgments. We further release a web app for visualization of Auto-ARGUE outputs.
Authors: Minhui Zhu, Minyang Tian, Xiaocheng Yang, Tianci Zhou, Penghao Zhu, Eli Chertkov, Shengyan Liu, Yufeng Du, Lifan Yuan, Ziming Ji, Indranil Das, Junyi Cao, Yufeng Du, Jinchen He, Yifan Su, Jiabin Yu, Yikun Jiang, Yujie Zhang, Chang Liu, Ze-Min Huang, Weizhen Jia, Xinan Chen, Peixue Wu, Yunkai Wang, Juntai Zhou, Yong Zhao, Farshid Jafarpour, Jessie Shelton, Aaron Young, John Bartolotta, Wenchao Xu, Yue Sun, Anjun Chu, Victor Colussi, Chris Akers, Nathan Brooks, Wenbo Fu, Christopher Wilson, Jinchao Zhao, Marvin Qi, Anqi Mu, Yubo Yang, Allen Zang, Yang Lyu, Peizhi Mai, Xuefei Guo, Luyu Gao, Ze Yang, Chi Xue, Dmytro Bandak, Ya\"ir Hein, Yonatan Kahn, Kevin Zhou, John Drew Wilson, Jarrod T. Reilly, Di Luo, Daniel Inafuku, Hao Tong, Liang Yang, Ruixing Zhang, Xueying Wang, Ofir Press, Nicolas Chia, Eliu Huerta, Hao Peng
Abstract: While large language models (LLMs) with reasoning capabilities are progressing rapidly on high-school math competitions and coding, can they reason effectively through complex, open-ended challenges found in frontier physics research? And crucially, what kinds of reasoning tasks do physicists want LLMs to assist with? To address these questions, we present the CritPt (Complex Research using Integrated Thinking - Physics Test, pronounced "critical point"), the first benchmark designed to test LLMs on unpublished, research-level reasoning tasks that broadly covers modern physics research areas, including condensed matter, quantum physics, atomic, molecular & optical physics, astrophysics, high energy physics, mathematical physics, statistical physics, nuclear physics, nonlinear dynamics, fluid dynamics and biophysics. CritPt consists of 71 composite research challenges designed to simulate full-scale research projects at the entry level, which are also decomposed to 190 simpler checkpoint tasks for more fine-grained insights. All problems are newly created by 50+ active physics researchers based on their own research. Every problem is hand-curated to admit a guess-resistant and machine-verifiable answer and is evaluated by an automated grading pipeline heavily customized for advanced physics-specific output formats. We find that while current state-of-the-art LLMs show early promise on isolated checkpoints, they remain far from being able to reliably solve full research-scale challenges: the best average accuracy among base models is only 4.0% , achieved by GPT-5 (high), moderately rising to around 10% when equipped with coding tools. Through the realistic yet standardized evaluation offered by CritPt, we highlight a large disconnect between current model capabilities and realistic physics research demands, offering a foundation to guide the development of scientifically grounded AI tools.