Authors: Harishwar Reddy, Madhusudan Srinivasan, Upulee Kanewala
Abstract: Large Language Models (LLMs) have made significant strides in Natural Language Processing but remain vulnerable to fairness-related issues, often reflecting biases inherent in their training data. These biases pose risks, particularly when LLMs are deployed in sensitive areas such as healthcare, finance, and law. This paper introduces a metamorphic testing approach to systematically identify fairness bugs in LLMs. We define and apply a set of fairness-oriented metamorphic relations (MRs) to assess the LLaMA and GPT model, a state-of-the-art LLM, across diverse demographic inputs. Our methodology includes generating source and follow-up test cases for each MR and analyzing model responses for fairness violations. The results demonstrate the effectiveness of MT in exposing bias patterns, especially in relation to tone and sentiment, and highlight specific intersections of sensitive attributes that frequently reveal fairness faults. This research improves fairness testing in LLMs, providing a structured approach to detect and mitigate biases and improve model robustness in fairness-sensitive applications.
Authors: Shurui Wu, Xinyi Huang, Dingxin Lu
Abstract: As the prevalence of mental health crises increases on social media platforms, identifying and preventing potential harm has become an urgent challenge. This study introduces a large language model (LLM)-based text transfer recognition method for social network crisis intervention, enhanced with domain-specific mental health knowledge. We propose a multi-level framework that incorporates transfer learning using BERT, and integrates mental health knowledge, sentiment analysis, and behavior prediction techniques. The framework includes a crisis annotation tool trained on social media datasets from real-world events, enabling the model to detect nuanced emotional cues and identify psychological crises. Experimental results show that the proposed method outperforms traditional models in crisis detection accuracy and exhibits greater sensitivity to subtle emotional and contextual variations.
Authors: Jianheng Li, Lirong Chen
Abstract: Research background: With the continuous development of society, consumers pay more attention to the key information of product fine-grained attributes when shopping. Research purposes: This study will fine tune the Sentence-BERT word embedding model and LDA model, mine the subject characteristics in online reviews of goods, and show consumers the details of various aspects of goods. Research methods: First, the Sentence-BERT model was fine tuned in the field of e-commerce online reviews, and the online review text was converted into a word vector set with richer semantic information; Secondly, the vectorized word set is input into the LDA model for topic feature extraction; Finally, focus on the key functions of the product through keyword analysis under the theme. Results: This study compared this model with other word embedding models and LDA models, and compared it with common topic extraction methods. The theme consistency of this model is 0.5 higher than that of other models, which improves the accuracy of theme extraction
Authors: Runjin Chen, Zhenyu Zhang, Junyuan Hong, Souvik Kundu, Zhangyang Wang
Abstract: Large Language Models (LLMs), such as OpenAI's o1-series have demonstrated compelling capabilities for complex reasoning tasks via the extended chain-of-thought (CoT) reasoning mechanism. However, recent studies reveal substantial redundancy in the CoT reasoning traces, which not only increases inference latency but also negatively impacts model performance by diverting attention to unnecessary reasoning paths. To address this issue, we investigate the internal reasoning structures of LLMs and categorize them into three primary thought types: execution, reflection, and transition thoughts. Moreover, our analysis reveals that excessive reflection and transition thoughts are strongly correlated with failure cases and these thought categories exhibit clear separation in the latent space. Based on these, we introduce SEAL (Steerable reasoning calibration), a training-free approach that seamlessly calibrates the CoT process, improving accuracy while demonstrating significant efficiency gains. SEAL consists of an offline stage for extracting the reasoning steering vector in the latent space, followed by an on-the-fly calibration of the reasoning trace through representation intervention using the steering vector. Notably, the steering vector exhibits strong transferability across various tasks. Extensive experiments across multiple models (DeepSeek-R1-Distill and QwQ-32B-Preview) and benchmarks (Math500, GSM8K, LiveCodeBench) validate the effectiveness of SEAL, up to a 11% improvement in accuracy while reducing reasoning tokens by 11.8% to 50.4%. Our code is publicly available at https://github.com/VITA-Group/SEAL.
Authors: Nirvan Patil, Malhar Abhay Inamdar, Agnivo Gosai, Guruprasad Pathak, Anish Joshi, Aryan Sagavekar, Anish Joshirao, Raj Dandekar, Rajat Dandekar, Sreedath Panat
Abstract: Small Language Models (SLMs) offer efficient alternatives to LLMs for specific domains. The 2023 TinyStories study developed an English dataset that allows SLMs with 1 to 10 million parameters to produce coherent outputs. Our research expands this framework by translating the original dataset into Indian languages and creating synthetic data using LLMs. We focus on Hindi, Marathi, and Bengali, evaluating SLMs for regional language processing and understanding linguistic complexity. We show that SLMs efficiently process regional languages with significantly fewer parameters than LLMs, providing a complementary framework for ``inference based evaluation" of tokenization strategies and linguistic complexity. Our analysis shows that language-specific tokenizers outperform general-purpose ones for Indian languages. Empirical validations, supported by information-theoretic and morphological analyses, provides fundamental understanding behind the better performance of Hindi models over Marathi and Bengali. Additionally, we show that synthetic datasets outperform translated content for training SLMs. Correlation analyses reveal cross-linguistic patterns and language-specific relationships between creativity, grammatical precision, and narrative completeness. These findings advance both the practical application of SLMs to underserved languages and our theoretical understanding of neural language development.
Authors: Seth Drake
Abstract: Large language model (LLM)-driven AI systems may exhibit an inference failure mode we term `neural howlround,' a self-reinforcing cognitive loop where certain highly weighted inputs become dominant, leading to entrenched response patterns resistant to correction. This paper explores the mechanisms underlying this phenomenon, which is distinct from model collapse and biased salience weighting. We propose an attenuation-based correction mechanism that dynamically introduces counterbalancing adjustments and can restore adaptive reasoning, even in `locked-in' AI systems. Additionally, we discuss some other related effects arising from improperly managed reinforcement. Finally, we outline potential applications of this mitigation strategy for improving AI robustness in real-world decision-making tasks.
Authors: Samah Alkhuzaey, Floriana Grasso, Terry R. Payne, Valentina Tamma
Abstract: Ontology-based question generation is an important application of semantic-aware systems that enables the creation of large question banks for diverse learning environments. The effectiveness of these systems, both in terms of the calibre and cognitive difficulty of the resulting questions, depends heavily on the quality and modelling approach of the underlying ontologies, making it crucial to assess their fitness for this task. To date, there has been no comprehensive investigation into the specific ontology aspects or characteristics that affect the question generation process. Therefore, this paper proposes a set of requirements and task-specific metrics for evaluating the fitness of ontologies for question generation tasks in pedagogical settings. Using the ROMEO methodology, a structured framework for deriving task-specific metrics, an expert-based approach is employed to assess the performance of various ontologies in Automatic Question Generation (AQG) tasks, which is then evaluated over a set of ontologies. Our results demonstrate that ontology characteristics significantly impact the effectiveness of question generation, with different ontologies exhibiting varying performance levels. This highlights the importance of assessing ontology quality with respect to AQG tasks.
Authors: Biplav Srivastava, Kausik Lakkaraju, Nitin Gupta, Vansh Nagpal, Bharath C. Muppasani, Sara E. Jones
Abstract: Collaborative assistants, or chatbots, are data-driven decision support systems that enable natural interaction for task completion. While they can meet critical needs in modern society, concerns about their reliability and trustworthiness persist. In particular, Large Language Model (LLM)-based chatbots like ChatGPT, Gemini, and DeepSeek are becoming more accessible. However, such chatbots have limitations, including their inability to explain response generation, the risk of generating problematic content, the lack of standardized testing for reliability, and the need for deep AI expertise and extended development times. These issues make chatbots unsuitable for trust-sensitive applications like elections or healthcare. To address these concerns, we introduce SafeChat, a general architecture for building safe and trustworthy chatbots, with a focus on information retrieval use cases. Key features of SafeChat include: (a) safety, with a domain-agnostic design where responses are grounded and traceable to approved sources (provenance), and 'do-not-respond' strategies to prevent harmful answers; (b) usability, with automatic extractive summarization of long responses, traceable to their sources, and automated trust assessments to communicate expected chatbot behavior, such as sentiment; and (c) fast, scalable development, including a CSV-driven workflow, automated testing, and integration with various devices. We implemented SafeChat in an executable framework using the open-source chatbot platform Rasa. A case study demonstrates its application in building ElectionBot-SC, a chatbot designed to safely disseminate official election information. SafeChat is being used in many domains, validating its potential, and is available at: https://github.com/ai4society/trustworthy-chatbot.
Authors: Tian Xie, Tongxin Yin, Vaishakh Keshava, Xueru Zhang, Siddhartha Reddy Jonnalagadda
Abstract: While large language models (LLMs) already play significant roles in society, research has shown that LLMs still generate content including social bias against certain sensitive groups. While existing benchmarks have effectively identified social biases in LLMs, a critical gap remains in our understanding of the underlying reasoning that leads to these biased outputs. This paper goes one step further to evaluate the causal reasoning process of LLMs when they answer questions eliciting social biases. We first propose a novel conceptual framework to classify the causal reasoning produced by LLMs. Next, we use LLMs to synthesize $1788$ questions covering $8$ sensitive attributes and manually validate them. The questions can test different kinds of causal reasoning by letting LLMs disclose their reasoning process with causal graphs. We then test 4 state-of-the-art LLMs. All models answer the majority of questions with biased causal reasoning, resulting in a total of $4135$ biased causal graphs. Meanwhile, we discover $3$ strategies for LLMs to avoid biased causal reasoning by analyzing the "bias-free" cases. Finally, we reveal that LLMs are also prone to "mistaken-biased" causal reasoning, where they first confuse correlation with causality to infer specific sensitive group names and then incorporate biased causal reasoning.
Authors: Miguel L\'opez-Otal, Jorge Gracia, Jordi Bernad, Carlos Bobed, Luc\'ia Pitarch-Ballesteros, Emma Angl\'es-Herrero
Abstract: Language models based on the Transformer architecture achieve excellent results in many language-related tasks, such as text classification or sentiment analysis. However, despite the architecture of these models being well-defined, little is known about how their internal computations help them achieve their results. This renders these models, as of today, a type of 'black box' systems. There is, however, a line of research -- 'interpretability' -- aiming to learn how information is encoded inside these models. More specifically, there is work dedicated to studying whether Transformer-based models possess knowledge of linguistic phenomena similar to human speakers -- an area we call 'linguistic interpretability' of these models. In this survey we present a comprehensive analysis of 160 research works, spread across multiple languages and models -- including multilingual ones -- that attempt to discover linguistic information from the perspective of several traditional Linguistics disciplines: Syntax, Morphology, Lexico-Semantics and Discourse. Our survey fills a gap in the existing interpretability literature, which either not focus on linguistic knowledge in these models or present some limitations -- e.g. only studying English-based models. Our survey also focuses on Pre-trained Language Models not further specialized for a downstream task, with an emphasis on works that use interpretability techniques that explore models' internal representations.
Authors: Tong Piao, Pei Tang, Zhipeng Zhang, Jiaqi Li, Qiao Liu, Zufeng Wu
Abstract: In recent years, Large Language Models (LLMs) have been widely applied across various domains due to their powerful domain adaptation capabilities. Previous studies have suggested that diverse, multi-modal data can enhance LLMs' domain adaptation performance. However, this hypothesis remains insufficiently validated in the e-commerce sector. To address this gap, we propose a comprehensive e-commerce multi-task framework and design empirical experiments to examine the impact of diverse data and tasks on LLMs from two perspectives: "capability comprehensiveness" and "task comprehensiveness." Specifically, we observe significant improvements in LLM performance by progressively introducing tasks related to new major capability areas and by continuously adding subtasks within different major capability domains. Furthermore, we observe that increasing model capacity amplifies the benefits of diversity, suggesting a synergistic relationship between model capacity and data diversity. Finally, we validate the best-performing model from our empirical experiments in the KDD Cup 2024, achieving a rank 5 in Task 1. This outcome demonstrates the significance of our research for advancing LLMs in the e-commerce domain.
Authors: Fabian Retkowski, Maike Z\"ufle, Andreas Sudmann, Dinah Pfau, Jan Niehues, Alexander Waibel
Abstract: Speech summarization has become an essential tool for efficiently managing and accessing the growing volume of spoken and audiovisual content. However, despite its increasing importance, speech summarization is still not clearly defined and intersects with several research areas, including speech recognition, text summarization, and specific applications like meeting summarization. This survey not only examines existing datasets and evaluation methodologies, which are crucial for assessing the effectiveness of summarization approaches but also synthesizes recent developments in the field, highlighting the shift from traditional systems to advanced models like fine-tuned cascaded architectures and end-to-end solutions.
Authors: Akram Mustafa, Usman Naseem, Mostafa Rahimi Azghadi
Abstract: Clinical document classification is essential for converting unstructured medical texts into standardised ICD-10 diagnoses, yet it faces challenges due to complex medical language, privacy constraints, and limited annotated datasets. Large Language Models (LLMs) offer promising improvements in accuracy and efficiency for this task. This study evaluates the performance and consistency of eight LLMs; four reasoning (Qwen QWQ, Deepseek Reasoner, GPT o3 Mini, Gemini 2.0 Flash Thinking) and four non-reasoning (Llama 3.3, GPT 4o Mini, Gemini 2.0 Flash, Deepseek Chat); in classifying clinical discharge summaries using the MIMIC-IV dataset. Using cTAKES to structure clinical narratives, models were assessed across three experimental runs, with majority voting determining final predictions. Results showed that reasoning models outperformed non-reasoning models in accuracy (71% vs 68%) and F1 score (67% vs 60%), with Gemini 2.0 Flash Thinking achieving the highest accuracy (75%) and F1 score (76%). However, non-reasoning models demonstrated greater stability (91% vs 84% consistency). Performance varied across ICD-10 codes, with reasoning models excelling in complex cases but struggling with abstract categories. Findings indicate a trade-off between accuracy and consistency, suggesting that a hybrid approach could optimise clinical coding. Future research should explore multi-label classification, domain-specific fine-tuning, and ensemble methods to enhance model reliability in real-world applications.
Authors: Ingryd V. S. T. Pereira, George D. C. Cavalcanti, Rafael M. O. Cruz
Abstract: Given the volume and speed at which fake news spreads across social media, automatic fake news detection has become a highly important task. However, this task presents several challenges, including extracting textual features that contain relevant information about fake news. Research about fake news detection shows that no single feature extraction technique consistently outperforms the others across all scenarios. Nevertheless, different feature extraction techniques can provide complementary information about the textual data and enable a more comprehensive representation of the content. This paper proposes using multi-view autoencoders to generate a joint feature representation for fake news detection by integrating several feature extraction techniques commonly used in the literature. Experiments on fake news datasets show a significant improvement in classification performance compared to individual views (feature representations). We also observed that selecting a subset of the views instead of composing a latent space with all the views can be advantageous in terms of accuracy and computational effort. For further details, including source codes, figures, and datasets, please refer to the project's repository: https://github.com/ingrydpereira/multiview-fake-news.
Authors: Daniil Larionov, Sotaro Takeshita, Ran Zhang, Yanran Chen, Christoph Leiter, Zhipin Wang, Christian Greisinger, Steffen Eger
Abstract: Reasoning-enabled large language models (LLMs) have recently demonstrated impressive performance in complex logical and mathematical tasks, yet their effectiveness in evaluating natural language generation remains unexplored. This study systematically compares reasoning-based LLMs (DeepSeek-R1 and OpenAI o3) with their non-reasoning counterparts across machine translation (MT) and text summarization (TS) evaluation tasks. We evaluate eight models across three architectural categories, including state-of-the-art reasoning models, their distilled variants (ranging from 8B to 70B parameters), and equivalent conventional, non-reasoning LLMs. Our experiments on WMT23 and SummEval benchmarks reveal that the benefits of reasoning capabilities are highly model and task-dependent: while OpenAI o3-mini models show consistent performance improvements with increased reasoning intensity, DeepSeek-R1 underperforms compared to its non-reasoning variant, with exception to certain aspects of TS evaluation. Correlation analysis demonstrates that increased reasoning token usage positively correlates with evaluation quality in o3-mini models. Furthermore, our results show that distillation of reasoning capabilities maintains reasonable performance in medium-sized models (32B) but degrades substantially in smaller variants (8B). This work provides the first comprehensive assessment of reasoning LLMs for NLG evaluation and offers insights into their practical use.
Authors: Alex Warstadt, Aaron Mueller, Leshem Choshen, Ethan Wilcox, Chengxu Zhuang, Juan Ciro, Rafael Mosquera, Bhargavi Paranjape, Adina Williams, Tal Linzen, Ryan Cotterell
Abstract: Children can acquire language from less than 100 million words of input. Large language models are far less data-efficient: they typically require 3 or 4 orders of magnitude more data and still do not perform as well as humans on many evaluations. These intensive resource demands limit the ability of researchers to train new models and use existing models as developmentally plausible cognitive models. The BabyLM Challenge is a communal effort in which participants compete to optimize language model training on a fixed data budget. Submissions are compared on various evaluation tasks targeting grammatical ability, downstream task performance, and generalization. Participants can submit to up to three tracks with progressively looser data restrictions. From over 30 submissions, we extract concrete recommendations on how best to train data-efficient language models, and on where future efforts should (and perhaps should not) focus. The winning submissions using the LTG-BERT architecture (Samuel et al., 2023) outperformed models trained on trillions of words. Other submissions achieved strong results through training on shorter input sequences or training a student model on a pretrained teacher. Curriculum learning attempts, which accounted for a large number of submissions, were largely unsuccessful, though some showed modest improvements.
Authors: Yu Fu, Haz Sameen Shahgir, Hui Liu, Xianfeng Tang, Qi He, Yue Dong
Abstract: Recent advances in long-context models (LCMs), designed to handle extremely long input contexts, primarily focus on utilizing external contextual information, often leaving the influence of large language models' intrinsic knowledge underexplored. In this work, we investigate how this intrinsic knowledge affects content generation and demonstrate that its impact becomes increasingly pronounced as context length extends. Furthermore, we show that the model's ability to utilize intrinsic knowledge, which we call intrinsic retrieval ability, does not improve simultaneously with its ability to leverage contextual knowledge through extrinsic retrieval ability. Moreover, better extrinsic retrieval can interfere with the model's ability to use its own knowledge effectively, limiting its full potential. To bridge this gap, we design a simple yet effective Hybrid Needle-in-a-Haystack test that evaluates models based on their capabilities across both retrieval abilities, rather than solely emphasizing extrinsic retrieval ability. Our experimental results reveal that Qwen-2.5 models significantly outperform Llama-3.1 models, demonstrating superior intrinsic retrieval ability. Moreover, even the more powerful Llama-3.1-70B-Instruct model fails to exhibit better performance under LCM conditions, highlighting the importance of evaluating models from a dual-retrieval perspective.
Authors: Leo Kampen, Carlos Rabat Villarreal, Louis Yu, Santu Karmaker, Dongji Feng
Abstract: In this paper, we conducted a Multi-Perspective Comparative Narrative Analysis (CNA) on three prominent LLMs: GPT-3.5, PaLM2, and Llama2. We applied identical prompts and evaluated their outputs on specific tasks, ensuring an equitable and unbiased comparison between various LLMs. Our study revealed that the three LLMs generated divergent responses to the same prompt, indicating notable discrepancies in their ability to comprehend and analyze the given task. Human evaluation was used as the gold standard, evaluating four perspectives to analyze differences in LLM performance.
Authors: Samuel Flanders, Melati Nungsari, Mark Cheong Wing Loong
Abstract: This study introduces a framework that leverages AI-generated descriptive codes to indicate a text's fecundity--the density of unique human-generated codes--in thematic analysis. Rather than replacing human interpretation, AI-generated codes guide the selection of texts likely to yield richer qualitative insights. Using a dataset of 2,530 Malaysian news articles on refugee attitudes, we compare AI-selected documents to randomly chosen ones by having three human coders independently derive codes. The results demonstrate that AI-selected texts exhibit approximately twice the fecundity. Our findings support the use of AI-generated codes as an effective proxy for identifying documents with a high potential for meaning-making in thematic analysis.
Authors: Tianyu Cao, Neel Bhandari, Akhila Yerukola, Akari Asai, Maarten Sap
Abstract: Despite the impressive performance of Retrieval-augmented Generation (RAG) systems across various NLP benchmarks, their robustness in handling real-world user-LLM interaction queries remains largely underexplored. This presents a critical gap for practical deployment, where user queries exhibit greater linguistic variations and can trigger cascading errors across interdependent RAG components. In this work, we systematically analyze how varying four linguistic dimensions (formality, readability, politeness, and grammatical correctness) impact RAG performance. We evaluate two retrieval models and nine LLMs, ranging from 3 to 72 billion parameters, across four information-seeking Question Answering (QA) datasets. Our results reveal that linguistic reformulations significantly impact both retrieval and generation stages, leading to a relative performance drop of up to 40.41% in Recall@5 scores for less formal queries and 38.86% in answer match scores for queries containing grammatical errors. Notably, RAG systems exhibit greater sensitivity to such variations compared to LLM-only generations, highlighting their vulnerability to error propagation due to linguistic shifts. These findings highlight the need for improved robustness techniques to enhance reliability in diverse user interactions.
Authors: Yonchanok Khaokaew, Flora D. Salim, Andreas Z\"ufle, Hao Xue, Taylor Anderson, Matthew Scotch, David J Heslop
Abstract: Generative agents have been increasingly used to simulate human behaviour in silico, driven by large language models (LLMs). These simulacra serve as sandboxes for studying human behaviour without compromising privacy or safety. However, it remains unclear whether such agents can truly represent real individuals. This work compares survey data from the Understanding America Study (UAS) on healthcare decision-making with simulated responses from generative agents. Using demographic-based prompt engineering, we create digital twins of survey respondents and analyse how well different LLMs reproduce real-world behaviours. Our findings show that some LLMs fail to reflect realistic decision-making, such as predicting universal vaccine acceptance. However, Llama 3 captures variations across race and Income more accurately but also introduces biases not present in the UAS data. This study highlights the potential of generative agents for behavioural research while underscoring the risks of bias from both LLMs and prompting strategies.
Authors: Vishal Gandhi, Sagar Gandhi
Abstract: Advancements in emotion aware language processing increasingly shape vital NLP applications ranging from conversational AI and affective computing to computational psychology and creative content generation. Existing emotion datasets either lack emotional granularity or fail to capture necessary stylistic diversity, limiting the advancement of effective emotion conditioned text generation systems. Seeking to bridge this crucial gap between granularity and style diversity, this paper introduces a novel systematically constructed dataset named ELSA Emotion and Language Style Alignment Dataset leveraging fine grained emotion taxonomies adapted from existing sources such as dair ai emotion dataset and GoEmotions taxonomy. This dataset comprises multiple emotionally nuanced variations of original sentences regenerated across distinct contextual styles such as conversational, formal, poetic, and narrative, using advanced Large Language Models LLMs. Rigorous computational evaluation using metrics such as perplexity, embedding variance, readability, lexical diversity, and semantic coherence measures validates the datasets emotional authenticity, linguistic fluency, and textual diversity. Comprehensive metric analyses affirm its potential to support deeper explorations into emotion conditioned style adaptive text generation. By enabling precision tuned emotionally nuanced language modeling, our dataset creates fertile ground for research on fine grained emotional control, prompt driven explanation, interpretability, and style adaptive expressive language generation with LLMs.
Authors: Yuyang Xu, Renjun Hu, Haochao Ying, Jian Wu, Xing Shi, Wei Lin
Abstract: Multiple-choice question (MCQ) benchmarks are widely used for evaluating Large Language Models (LLMs), yet their reliability is undermined by benchmark contamination. In this study, we reframe contamination as an inherent aspect of learning and seek to disentangle genuine capability acquisition from superficial memorization in LLM evaluation. First, by analyzing model performance under different memorization conditions, we uncover a counterintuitive trend: LLMs perform worse on memorized MCQs than on non-memorized ones, indicating the coexistence of two distinct learning phenomena, i.e., rote memorization and genuine capability learning. To disentangle them, we propose TrinEval, a novel evaluation framework that reformulates MCQs into an alternative trinity format, reducing memorization while preserving knowledge assessment. Experiments validate TrinEval's effectiveness in reformulation, and its evaluation reveals that common LLMs may memorize by rote 20.5% of knowledge points (in MMLU on average).
Authors: Markus Flicke, Glenn Angrabeit, Madhav Iyengar, Vitalii Protsenko, Illia Shakun, Jovan Cicvaric, Bora Kargi, Haoyu He, Lukas Schuler, Lewin Scholz, Kavyanjali Agnihotri, Yong Cao, Andreas Geiger
Abstract: Scholar Inbox is a new open-access platform designed to address the challenges researchers face in staying current with the rapidly expanding volume of scientific literature. We provide personalized recommendations, continuous updates from open-access archives (arXiv, bioRxiv, etc.), visual paper summaries, semantic search, and a range of tools to streamline research workflows and promote open research access. The platform's personalized recommendation system is trained on user ratings, ensuring that recommendations are tailored to individual researchers' interests. To further enhance the user experience, Scholar Inbox also offers a map of science that provides an overview of research across domains, enabling users to easily explore specific topics. We use this map to address the cold start problem common in recommender systems, as well as an active learning strategy that iteratively prompts users to rate a selection of papers, allowing the system to learn user preferences quickly. We evaluate the quality of our recommendation system on a novel dataset of 800k user ratings, which we make publicly available, as well as via an extensive user study. https://www.scholar-inbox.com/
Authors: Yin Jou Huang, Rafik Hadfi
Abstract: There is a growing interest in assessing the personality traits of Large language models (LLMs). However, traditional personality assessments based on self-report questionnaires may fail to capture their true behavioral nuances due to inherent biases and meta-knowledge contamination. This paper introduces a novel multi-observer framework for LLM personality assessment that draws inspiration from informant-report methods in psychology. Instead of relying solely on self-assessments, our approach employs multiple observer agents configured with a specific relationship context (e.g., family, friend, or workplace) to simulate interactive scenarios with a subject LLM. These observers engage in dialogues and subsequently provide ratings across the Big Five personality dimensions. Our experiments reveal that LLMs possess systematic biases in self-report personality ratings. Moreover, aggregating observer ratings effectively reduces non-systematic biases and achieves optimal reliability with 5-7 observers. The findings highlight the significant impact of relationship context on personality perception and demonstrate that a multi-observer paradigm yields a more robust and context-sensitive evaluation of LLM personality traits.
Authors: Taisei Kanda, Mingzhe Jin, Wataru Zaitsu
Abstract: Traditionally, authorship attribution (AA) tasks relied on statistical data analysis and classification based on stylistic features extracted from texts. In recent years, pre-trained language models (PLMs) have attracted significant attention in text classification tasks. However, although they demonstrate excellent performance on large-scale short-text datasets, their effectiveness remains under-explored for small samples, particularly in AA tasks. Additionally, a key challenge is how to effectively leverage PLMs in conjunction with traditional feature-based methods to advance AA research. In this study, we aimed to significantly improve performance using an integrated integrative ensemble of traditional feature-based and modern PLM-based methods on an AA task in a small sample. For the experiment, we used two corpora of literary works to classify 10 authors each. The results indicate that BERT is effective, even for small-sample AA tasks. Both BERT-based and classifier ensembles outperformed their respective stand-alone models, and the integrated ensemble approach further improved the scores significantly. For the corpus that was not included in the pre-training data, the integrated ensemble improved the F1 score by approximately 14 points, compared to the best-performing single model. Our methodology provides a viable solution for the efficient use of the ever-expanding array of data processing tools in the foreseeable future.
Authors: Siddhant Arora, Kai-Wei Chang, Chung-Ming Chien, Yifan Peng, Haibin Wu, Yossi Adi, Emmanuel Dupoux, Hung-Yi Lee, Karen Livescu, Shinji Watanabe
Abstract: The field of spoken language processing is undergoing a shift from training custom-built, task-specific models toward using and optimizing spoken language models (SLMs) which act as universal speech processing systems. This trend is similar to the progression toward universal language models that has taken place in the field of (text) natural language processing. SLMs include both "pure" language models of speech -- models of the distribution of tokenized speech sequences -- and models that combine speech encoders with text language models, often including both spoken and written input or output. Work in this area is very diverse, with a range of terminology and evaluation settings. This paper aims to contribute an improved understanding of SLMs via a unifying literature survey of recent work in the context of the evolution of the field. Our survey categorizes the work in this area by model architecture, training, and evaluation choices, and describes some key challenges and directions for future work.
Authors: Burak Senel
Abstract: Automated scoring (AS) systems are increasingly used for evaluating L2 writing, but require ongoing refinement for construct validity. While prior work suggested lexical bundles (LBs) - recurrent multi-word sequences satisfying certain frequency criteria - could inform assessment, their empirical integration into AS models needs further investigation. This study tested the impact of incorporating LB frequency features into an AS model for TOEFL independent writing tasks. Analyzing a sampled subcorpus (N=1,225 essays, 9 L1s) from the TOEFL11 corpus, scored by ETS-trained raters (Low, Medium, High), 3- to 9-word LBs were extracted, distinguishing prompt-specific from non-prompt types. A baseline Support Vector Machine (SVM) scoring model using established linguistic features (e.g., mechanics, cohesion, sophistication) was compared against an extended model including three aggregate LB frequency features (total prompt, total non-prompt, overall total). Results revealed significant, though generally small-effect, relationships between LB frequency (especially non-prompt bundles) and proficiency (p < .05). Mean frequencies suggested lower proficiency essays used more LBs overall. Critically, the LB-enhanced model improved agreement with human raters (Quadratic Cohen's Kappa +2.05%, overall Cohen's Kappa +5.63%), with notable gains for low (+10.1% exact agreement) and medium (+14.3% Cohen's Kappa) proficiency essays. These findings demonstrate that integrating aggregate LB frequency offers potential for developing more linguistically informed and accurate AS systems, particularly for differentiating developing L2 writers.
Authors: Frances Laureano De Leon, Yixiao Wang, Yue Feng, Mark G. Lee
Abstract: Emotion detection in natural language processing is a challenging task due to the complexity of human emotions and linguistic diversity. While significant progress has been made in high-resource languages, emotion detection in low-resource languages remains underexplored. In this work, we address multilingual and cross-lingual emotion detection by leveraging adapter-based fine-tuning with multilingual pre-trained language models. Adapters introduce a small number of trainable parameters while keeping the pre-trained model weights fixed, offering a parameter-efficient approach to adaptation. We experiment with different adapter tuning strategies, including task-only adapters, target-language-ready task adapters, and language-family-based adapters. Our results show that target-language-ready task adapters achieve the best overall performance, particularly for low-resource African languages with our team ranking 7th for Tigrinya, and 8th for Kinyarwanda in Track A. In Track C, our system ranked 3rd for Amharic, and 4th for Oromo, Tigrinya, Kinyarwanda, Hausa, and Igbo. Our approach outperforms large language models in 11 languages and matches their performance in four others, despite our models having significantly fewer parameters. Furthermore, we find that adapter-based models retain cross-linguistic transfer capabilities while requiring fewer computational resources compared to full fine-tuning for each language.
Authors: Nicola Horst, Davide Mazzaccara, Antonia Schmidt, Michael Sullivan, Filippo Moment\`e, Luca Franceschetti, Philipp Sadler, Sherzod Hakimov, Alberto Testoni, Raffaella Bernardi, Raquel Fern\'andez, Alexander Koller, Oliver Lemon, David Schlangen, Mario Giulianelli, Alessandro Suglia
Abstract: Are we running out of learning signal? Predicting the next word in an existing text has turned out to be a powerful signal, at least at scale. But there are signs that we are running out of this resource. In recent months, interaction between learner and feedback-giver has come into focus, both for "alignment" (with a reward model judging the quality of instruction following attempts) and for improving "reasoning" (process- and outcome-based verifiers judging reasoning steps). In this paper, we explore to what extent synthetic interaction in what we call Dialogue Games -- goal-directed and rule-governed activities driven predominantly by verbal actions -- can provide a learning signal, and how this signal can be used. We introduce an environment for producing such interaction data (with the help of a Large Language Model as counterpart to the learner model), both offline and online. We investigate the effects of supervised fine-tuning on this data, as well as reinforcement learning setups such as DPO, and GRPO; showing that all of these approaches achieve some improvements in in-domain games, but only GRPO demonstrates the ability to generalise to out-of-domain games as well as retain competitive performance in reference-based tasks. We release the framework and the baseline training setups in the hope that this can foster research in this promising new direction.
Authors: Gaya Mehenni, Amal Zouaq
Abstract: We present MedHal, a novel large-scale dataset specifically designed to evaluate if models can detect hallucinations in medical texts. Current hallucination detection methods face significant limitations when applied to specialized domains like medicine, where they can have disastrous consequences. Existing medical datasets are either too small, containing only a few hundred samples, or focus on a single task like Question Answering or Natural Language Inference. MedHal addresses these gaps by: (1) incorporating diverse medical text sources and tasks; (2) providing a substantial volume of annotated samples suitable for training medical hallucination detection models; and (3) including explanations for factual inconsistencies to guide model learning. We demonstrate MedHal's utility by training and evaluating a baseline medical hallucination detection model, showing improvements over general-purpose hallucination detection approaches. This resource enables more efficient evaluation of medical text generation systems while reducing reliance on costly expert review, potentially accelerating the development of medical AI research.
Authors: Julian B\"aumler, Louis Bl\"ocher, Lars-Joel Frey, Xian Chen, Markus Bayer, Christian Reuter
Abstract: The dissemination of online hate speech can have serious negative consequences for individuals, online communities, and entire societies. This and the large volume of hateful online content prompted both practitioners', i.e., in content moderation or law enforcement, and researchers' interest in machine learning models to automatically classify instances of hate speech. Whereas most scientific works address hate speech classification as a binary task, practice often requires a differentiation into sub-types, e.g., according to target, severity, or legality, which may overlap for individual content. Hence, researchers created datasets and machine learning models that approach hate speech classification in textual data as a multi-label problem. This work presents the first systematic and comprehensive survey of scientific literature on this emerging research landscape in English (N=46). We contribute with a concise overview of 28 datasets suited for training multi-label classification models that reveals significant heterogeneity regarding label-set, size, meta-concept, annotation process, and inter-annotator agreement. Our analysis of 24 publications proposing suitable classification models further establishes inconsistency in evaluation and a preference for architectures based on Bidirectional Encoder Representation from Transformers (BERT) and Recurrent Neural Networks (RNNs). We identify imbalanced training data, reliance on crowdsourcing platforms, small and sparse datasets, and missing methodological alignment as critical open issues and formulate ten recommendations for research.
Authors: Fangzhi Xu, Hang Yan, Chang Ma, Haiteng Zhao, Qiushi Sun, Kanzhi Cheng, Junxian He, Jun Liu, Zhiyong Wu
Abstract: Advancing LLM reasoning skills has captivated wide interest. However, current post-training techniques rely heavily on supervisory signals, such as outcome supervision or auxiliary reward models, which face the problem of scalability and high annotation costs. This motivates us to enhance LLM reasoning without the need for external supervision. We introduce a generalizable and purely unsupervised self-training framework, named Genius. Without external auxiliary, Genius requires to seek the optimal response sequence in a stepwise manner and optimize the LLM. To explore the potential steps and exploit the optimal ones, Genius introduces a stepwise foresight re-sampling strategy to sample and estimate the step value by simulating future outcomes. Further, we recognize that the unsupervised setting inevitably induces the intrinsic noise and uncertainty. To provide a robust optimization, we propose an advantage-calibrated optimization (ACO) loss function to mitigate estimation inconsistencies. Combining these techniques together, Genius provides an advanced initial step towards self-improve LLM reasoning with general queries and without supervision, revolutionizing reasoning scaling laws given the vast availability of general queries. The code will be released at https://github.com/xufangzhi/Genius.
Authors: Yiliu Sun, Yanfang Zhang, Zicheng Zhao, Sheng Wan, Dacheng Tao, Chen Gong
Abstract: Nowadays, Large Language Models (LLMs) have been gradually employed to solve complex tasks. To face the challenge, task decomposition has become an effective way, which proposes to divide a complex task into multiple simpler subtasks and then solve them separately so that the difficulty of the original task can be reduced. However, the performance of existing task decomposition methods can be suboptimal when the task contains overly complex logic and constraints. In this situation, the solution generated by LLMs may deviate from the original purpose of the task, or contain redundant or even erroneous content. Therefore, inspired by the fact that humans possess two thinking systems including fast thinking and slow thinking, this paper introduces a new task decomposition method termed ``Fast-Slow-Thinking'' (FST), which stimulates LLMs to solve tasks through the cooperation of Fast Thinking (FT) and Slow Thinking (ST) steps. Here FT focuses more on the general and concise aspect of the task, and ST focuses more on the details of the task. In FT, LLMs are prompted to remove the constraints of the original task, therefore simplifying it to a general and concise one. In ST, we recall the constraints removed in FT, so that LLMs can improve the answer generated in FT to meet the requirements of the original task. Therefore, our FST method enables LLMs to consider a complex problem via a human-like cognition process from coarse to fine, the effectiveness of which has been well demonstrated by the experiments on three types of tasks.
Authors: Hang Ni, Fan Liu, Xinyu Ma, Lixin Su, Shuaiqiang Wang, Dawei Yin, Hui Xiong, Hao Liu
Abstract: Large language models (LLMs) have shown promise in automating travel planning, yet they often fall short in addressing nuanced spatiotemporal rationality. While existing benchmarks focus on basic plan validity, they neglect critical aspects such as route efficiency, POI appeal, and real-time adaptability. This paper introduces TP-RAG, the first benchmark tailored for retrieval-augmented, spatiotemporal-aware travel planning. Our dataset includes 2,348 real-world travel queries, 85,575 fine-grain annotated POIs, and 18,784 high-quality travel trajectory references sourced from online tourist documents, enabling dynamic and context-aware planning. Through extensive experiments, we reveal that integrating reference trajectories significantly improves spatial efficiency and POI rationality of the travel plan, while challenges persist in universality and robustness due to conflicting references and noisy data. To address these issues, we propose EvoRAG, an evolutionary framework that potently synergizes diverse retrieved trajectories with LLMs' intrinsic reasoning. EvoRAG achieves state-of-the-art performance, improving spatiotemporal compliance and reducing commonsense violation compared to ground-up and retrieval-augmented baselines. Our work underscores the potential of hybridizing Web knowledge with LLM-driven optimization, paving the way for more reliable and adaptive travel planning agents.
Authors: Zden\v{e}k Kasner, Vil\'em Zouhar, Patr\'icia Schmidtov\'a, Ivan Kart\'a\v{c}, Krist\'yna Onderkov\'a, Ond\v{r}ej Pl\'atek, Dimitra Gkatzia, Saad Mahamood, Ond\v{r}ej Du\v{s}ek, Simone Balloccu
Abstract: For high-quality texts, single-score metrics seldom provide actionable feedback. In contrast, span annotation - pointing out issues in the text by annotating their spans - can guide improvements and provide insights. Until recently, span annotation was limited to human annotators or fine-tuned encoder models. In this study, we automate span annotation with large language models (LLMs). We compare expert or skilled crowdworker annotators with open and proprietary LLMs on three tasks: data-to-text generation evaluation, machine translation evaluation, and propaganda detection in human-written texts. In our experiments, we show that LLMs as span annotators are straightforward to implement and notably more cost-efficient than human annotators. The LLMs achieve moderate agreement with skilled human annotators, in some scenarios comparable to the average agreement among the annotators themselves. Qualitative analysis shows that reasoning models outperform their instruction-tuned counterparts and provide more valid explanations for annotations. We release the dataset of more than 40k model and human annotations for further research.
Authors: Wissam Antoun, Beno\^it Sagot, Djam\'e Seddah
Abstract: Pretrained transformer-encoder models like DeBERTaV3 and ModernBERT introduce architectural advancements aimed at improving efficiency and performance. Although the authors of ModernBERT report improved performance over DeBERTaV3 on several benchmarks, the lack of disclosed training data and the absence of comparisons using a shared dataset make it difficult to determine whether these gains are due to architectural improvements or differences in training data. In this work, we conduct a controlled study by pretraining ModernBERT on the same dataset as CamemBERTaV2, a DeBERTaV3 French model, isolating the effect of model design. Our results show that the previous model generation remains superior in sample efficiency and overall benchmark performance, with ModernBERT's primary advantage being faster training and inference speed. However, the new proposed model still provides meaningful architectural improvements compared to earlier models such as BERT and RoBERTa. Additionally, we observe that high-quality pre-training data accelerates convergence but does not significantly improve final performance, suggesting potential benchmark saturation. These findings show the importance of disentangling pretraining data from architectural innovations when evaluating transformer models.
Authors: Krishna C. Puvvada, Faisal Ladhak, Santiago Akle Serrano, Cheng-Ping Hsieh, Shantanu Acharya, Somshubra Majumdar, Fei Jia, Samuel Kriman, Simeng Sun, Dima Rekesh, Boris Ginsburg
Abstract: We present a decoder-only Transformer architecture that robustly generalizes to sequence lengths substantially longer than those seen during training. Our model, SWAN-GPT, interleaves layers without positional encodings (NoPE) and sliding-window attention layers equipped with rotary positional encodings (SWA-RoPE). Experiments demonstrate strong performance on sequence lengths significantly longer than the training length without the need for additional long-context training. This robust length extrapolation is achieved through our novel architecture, enhanced by a straightforward dynamic scaling of attention scores during inference. In addition, SWAN-GPT is more computationally efficient than standard GPT architectures, resulting in cheaper training and higher throughput. Further, we demonstrate that existing pre-trained decoder-only models can be efficiently converted to the SWAN architecture with minimal continued training, enabling longer contexts. Overall, our work presents an effective approach for scaling language models to longer contexts in a robust and efficient manner.
Authors: Tanmay Laud, Akadia Kacha-Ochana, Steven A. Sumner, Vikram Krishnasamy, Royal Law, Lyna Schieber, Munmun De Choudhury, Mai ElSherief
Abstract: Opioid use disorder (OUD) is a leading health problem that affects individual well-being as well as general public health. Due to a variety of reasons, including the stigma faced by people using opioids, online communities for recovery and support were formed on different social media platforms. In these communities, people share their experiences and solicit information by asking questions to learn about opioid use and recovery. However, these communities do not always contain clinically verified information. In this paper, we study natural language questions asked in the context of OUD-related discourse on Reddit. We adopt transformer-based question detection along with hierarchical clustering across 19 subreddits to identify six coarse-grained categories and 69 fine-grained categories of OUD-related questions. Our analysis uncovers ten areas of information seeking from Reddit users in the context of OUD: drug sales, specific drug-related questions, OUD treatment, drug uses, side effects, withdrawal, lifestyle, drug testing, pain management and others, during the study period of 2018-2021. Our work provides a major step in improving the understanding of OUD-related questions people ask unobtrusively on Reddit. We finally discuss technological interventions and public health harm reduction techniques based on the topics of these questions.
Authors: Yutaro Yamada, Robert Tjarko Lange, Cong Lu, Shengran Hu, Chris Lu, Jakob Foerster, Jeff Clune, David Ha
Abstract: AI is increasingly playing a pivotal role in transforming how scientific discoveries are made. We introduce The AI Scientist-v2, an end-to-end agentic system capable of producing the first entirely AI generated peer-review-accepted workshop paper. This system iteratively formulates scientific hypotheses, designs and executes experiments, analyzes and visualizes data, and autonomously authors scientific manuscripts. Compared to its predecessor (v1, Lu et al., 2024 arXiv:2408.06292), The AI Scientist-v2 eliminates the reliance on human-authored code templates, generalizes effectively across diverse machine learning domains, and leverages a novel progressive agentic tree-search methodology managed by a dedicated experiment manager agent. Additionally, we enhance the AI reviewer component by integrating a Vision-Language Model (VLM) feedback loop for iterative refinement of content and aesthetics of the figures. We evaluated The AI Scientist-v2 by submitting three fully autonomous manuscripts to a peer-reviewed ICLR workshop. Notably, one manuscript achieved high enough scores to exceed the average human acceptance threshold, marking the first instance of a fully AI-generated paper successfully navigating a peer review. This accomplishment highlights the growing capability of AI in conducting all aspects of scientific research. We anticipate that further advancements in autonomous scientific discovery technologies will profoundly impact human knowledge generation, enabling unprecedented scalability in research productivity and significantly accelerating scientific breakthroughs, greatly benefiting society at large. We have open-sourced the code at https://github.com/SakanaAI/AI-Scientist-v2 to foster the future development of this transformative technology. We also discuss the role of AI in science, including AI safety.
Authors: Tianyi Wu, Zhiwei Xue, Yue Liu, Jiaheng Zhang, Bryan Hooi, See-Kiong Ng
Abstract: Jailbreak attacks, which aim to cause LLMs to perform unrestricted behaviors, have become a critical and challenging direction in AI safety. Despite achieving the promising attack success rate using dictionary-based evaluation, existing jailbreak attack methods fail to output detailed contents to satisfy the harmful request, leading to poor performance on GPT-based evaluation. To this end, we propose a black-box jailbreak attack termed GeneShift, by using a genetic algorithm to optimize the scenario shifts. Firstly, we observe that the malicious queries perform optimally under different scenario shifts. Based on it, we develop a genetic algorithm to evolve and select the hybrid of scenario shifts. It guides our method to elicit detailed and actionable harmful responses while keeping the seemingly benign facade, improving stealthiness. Extensive experiments demonstrate the superiority of GeneShift. Notably, GeneShift increases the jailbreak success rate from 0% to 60% when direct prompting alone would fail.
Authors: Aashiq Muhamed, Jacopo Bonato, Mona Diab, Virginia Smith
Abstract: Machine unlearning is a promising approach to improve LLM safety by removing unwanted knowledge from the model. However, prevailing gradient-based unlearning methods suffer from issues such as high computational costs, hyperparameter instability, poor sequential unlearning capability, vulnerability to relearning attacks, low data efficiency, and lack of interpretability. While Sparse Autoencoders are well-suited to improve these aspects by enabling targeted activation-based unlearning, prior approaches underperform gradient-based methods. This work demonstrates that, contrary to these earlier findings, SAEs can significantly improve unlearning when employed dynamically. We introduce $\textbf{Dynamic DAE Guardrails}$ (DSG), a novel method for precision unlearning that leverages principled feature selection and a dynamic classifier. Our experiments show DSG substantially outperforms leading unlearning methods, achieving superior forget-utility trade-offs. DSG addresses key drawbacks of gradient-based approaches for unlearning -- offering enhanced computational efficiency and stability, robust performance in sequential unlearning, stronger resistance to relearning attacks, better data efficiency including zero-shot settings, and more interpretable unlearning.
Authors: Liu Xiao, Li Zhiyuan, Lin Yueyu
Abstract: State-based sequence models like RWKV-7 offer a compelling alternative to Transformer architectures, achieving linear complexity while demonstrating greater expressive power in short-context scenarios and enabling state tracking beyond the \(\text{TC}^0\) complexity class. However, RWKV-7 lacks mechanisms for token-parameter interactions and native scalability, limiting its adaptability and growth without retraining. In this paper, we propose \textbf{Meta-State}, a novel extension to RWKV-7 that replaces attention mechanisms with a fully state-driven approach, integrating token-parameter interactions through a \textbf{Self-State Encoder} (SSE) mechanism. The SSE repurposes a portion of the RWKV-7 Weighted Key-Value (WKV) state as transformation weights to encode token-parameter interactions in a linear, state-driven manner without introducing new trainable matrices or softmax operations, while preserving the autoregressive property of token processing. Meta-State supports progressive model scaling by expanding the WKV state and parameter tokens, reusing existing parameters without retraining. Our approach bridges the gap between state-based modeling, token-parameter interactions, and scalable architectures, offering a flexible framework for efficient and adaptable sequence modeling with linear complexity and constant memory usage.
Authors: Qi Zhi Lim, Chin Poo Lee, Kian Ming Lim, Kalaiarasi Sonai Muthu Anbananthen
Abstract: The increasing availability of multimodal data across text, tables, and images presents new challenges for developing models capable of complex cross-modal reasoning. Existing methods for Multimodal Multi-hop Question Answering (MMQA) often suffer from limited reasoning capabilities, reliance on modality conversion, and inadequate alignment between visual and textual representations. To address these limitations, this paper introduces Vision-Language Multimodal Transformer (VLMT), a unified architecture that integrates a transformer-based vision encoder with a sequence-to-sequence language model. VLMT employs a direct token-level injection mechanism to fuse visual and textual inputs within a shared embedding space, eliminating the need for intermediate projection layers. To enhance cross-modal alignment and reasoning, a three-stage pretraining strategy is proposed to progressively align vision-language representations and improve the model's capacity for multimodal understanding. Based on the pretrained backbone, two task-specific modules are instantiated to form a two-stage MMQA framework: a multimodal reranker that predicts document relevance scores and utilizes a relative threshold with top-k strategy for context retrieval, and a multimodal question answering model that generates contextually grounded answers based on the retrieved evidence. Comprehensive experiments on two benchmark datasets demonstrate the effectiveness of the proposed approach. On MultimodalQA validation set, VLMT-Large achieves 76.5% Exact Match and 80.1% F1, outperforming the previous state-of-the-art by +9.1% in Exact Match and +8.8% in F1. On WebQA, it attains a QA score of 47.6, surpassing prior models such as PERQA by +3.2. These results highlight VLMT's strong capabilities in multimodal reasoning and its potential to advance real-world information retrieval and question answering systems.
Authors: Haowei Lou, Hye-young Paik, Sheng Li, Wen Hu, Lina Yao
Abstract: Text-to-Speech (TTS) models can generate natural, human-like speech across multiple languages by transforming phonemes into waveforms. However, multilingual TTS remains challenging due to discrepancies in phoneme vocabularies and variations in prosody and speaking style across languages. Existing approaches either train separate models for each language, which achieve high performance at the cost of increased computational resources, or use a unified model for multiple languages that struggles to capture fine-grained, language-specific style variations. In this work, we propose LanStyleTTS, a non-autoregressive, language-aware style adaptive TTS framework that standardizes phoneme representations and enables fine-grained, phoneme-level style control across languages. This design supports a unified multilingual TTS model capable of producing accurate and high-quality speech without the need to train language-specific models. We evaluate LanStyleTTS by integrating it with several state-of-the-art non-autoregressive TTS architectures. Results show consistent performance improvements across different model backbones. Furthermore, we investigate a range of acoustic feature representations, including mel-spectrograms and autoencoder-derived latent features. Our experiments demonstrate that latent encodings can significantly reduce model size and computational cost while preserving high-quality speech generation.
Authors: Junmo Kim, Namkyeong Lee, Jiwon Kim, Kwangsoo Kim
Abstract: Electronic health record (EHR) foundation models have been an area ripe for exploration with their improved performance in various medical tasks. Despite the rapid advances, there exists a fundamental limitation: Processing unseen medical codes out of the vocabulary. This problem limits the generality of EHR foundation models and the integration of models trained with different vocabularies. To deal with this problem, we propose MedRep for EHR foundation models based on the observational medical outcome partnership (OMOP) common data model (CDM), providing the integrated medical concept representations and the basic data augmentation strategy for patient trajectories. For concept representation learning, we enrich the information of each concept with a minimal definition through large language model (LLM) prompts and enhance the text-based representations through graph ontology of OMOP vocabulary. Trajectory augmentation randomly replaces selected concepts with other similar concepts that have closely related representations to let the model practice with the concepts out-of-vocabulary. Finally, we demonstrate that EHR foundation models trained with MedRep better maintain the prediction performance in external datasets. Our code implementation is publicly available at https://github.com/kicarussays/MedRep.
Authors: Cheng-Yu Hsieh, Pavan Kumar Anasosalu Vasu, Fartash Faghri, Raviteja Vemulapalli, Chun-Liang Li, Ranjay Krishna, Oncel Tuzel, Hadi Pouransari
Abstract: Visual understanding is inherently contextual -- what we focus on in an image depends on the task at hand. For instance, given an image of a person holding a bouquet of flowers, we may focus on either the person such as their clothing, or the type of flowers, depending on the context of interest. Yet, most existing image encoding paradigms represent an image as a fixed, generic feature vector, overlooking the potential needs of prioritizing varying visual information for different downstream use cases. In this work, we introduce FocalLens, a conditional visual encoding method that produces different representations for the same image based on the context of interest, expressed flexibly through natural language. We leverage vision instruction tuning data and contrastively finetune a pretrained vision encoder to take natural language instructions as additional inputs for producing conditional image representations. Extensive experiments validate that conditional image representation from FocalLens better pronounce the visual features of interest compared to generic features produced by standard vision encoders like CLIP. In addition, we show FocalLens further leads to performance improvements on a range of downstream tasks including image-image retrieval, image classification, and image-text retrieval, with an average gain of 5 and 10 points on the challenging SugarCrepe and MMVP-VLM benchmarks, respectively.
Authors: Md Abdullah Al Kafi, Sumit Kumar Banshal, Md Sadman Shakib, Showrov Azam, Tamanna Alam Tabashom
Abstract: One of the most alarming issues in digital society is hate speech (HS) on social media. The severity is so high that researchers across the globe are captivated by this domain. A notable amount of work has been conducted to address the identification and alarm system. However, a noticeable gap exists, especially for low-resource languages. Comprehensive datasets are the main problem among the constrained resource languages, such as Bangla. Interestingly, hate speech or any particular speech has no single dimensionality. Similarly, the hate component can simultaneously have multiple abusive attributes, which seems to be missed in the existing datasets. Thus, a multi-label Bangla hate speech dataset named BOISHOMMO has been compiled and evaluated in this work. That includes categories of HS across race, gender, religion, politics, and more. With over two thousand annotated examples, BOISHOMMO provides a nuanced understanding of hate speech in Bangla and highlights the complexities of processing non-Latin scripts. Apart from evaluating with multiple algorithmic approaches, it also highlights the complexities of processing Bangla text and assesses model performance. This unique multi-label approach enriches future hate speech detection and analysis studies for low-resource languages by providing a more nuanced, diverse dataset.
Authors: Ye Ye
Abstract: Large Language Models (LLMs) are increasingly used as autonomous agents for multi-step tasks. However, most existing frameworks fail to maintain a structured understanding of the task state, often relying on linear prompt concatenation or shallow memory buffers. This leads to brittle performance, frequent hallucinations, and poor long-range coherence. In this work, we propose the Task Memory Engine (TME), a lightweight and structured memory module that tracks task execution using a hierarchical Task Memory Tree (TMT). Each node in the tree corresponds to a task step, storing relevant input, output, status, and sub-task relationships. We introduce a prompt synthesis method that dynamically generates LLM prompts based on the active node path, significantly improving execution consistency and contextual grounding. Through case studies and comparative experiments on multi-step agent tasks, we demonstrate that TME leads to better task completion accuracy and more interpretable behavior with minimal implementation overhead. The full implementation of TME is available at https://github.com/biubiutomato/TME-Agent.
Authors: Zhiqiu Xia, Lang Zhu, Bingzhe Li, Feng Chen, Qiannan Li, Hang Liu
Abstract: Large Language Models (LLMs) are reshaping the landscape of computer science research, driving significant shifts in research priorities across diverse conferences and fields. This study provides a comprehensive analysis of the publication trend of LLM-related papers in 77 top-tier computer science conferences over the past six years (2019-2024). We approach this analysis from four distinct perspectives: (1) We investigate how LLM research is driving topic shifts within major conferences. (2) We adopt a topic modeling approach to identify various areas of LLM-related topic growth and reveal the topics of concern at different conferences. (3) We explore distinct contribution patterns of academic and industrial institutions. (4) We study the influence of national origins on LLM development trajectories. Synthesizing the findings from these diverse analytical angles, we derive ten key insights that illuminate the dynamics and evolution of the LLM research ecosystem.
Authors: Jialu Li, Shoubin Yu, Han Lin, Jaemin Cho, Jaehong Yoon, Mohit Bansal
Abstract: Recent advancements in text-to-video (T2V) diffusion models have significantly enhanced the visual quality of the generated videos. However, even recent T2V models find it challenging to follow text descriptions accurately, especially when the prompt requires accurate control of spatial layouts or object trajectories. A recent line of research uses layout guidance for T2V models that require fine-tuning or iterative manipulation of the attention map during inference time. This significantly increases the memory requirement, making it difficult to adopt a large T2V model as a backbone. To address this, we introduce Video-MSG, a training-free Guidance method for T2V generation based on Multimodal planning and Structured noise initialization. Video-MSG consists of three steps, where in the first two steps, Video-MSG creates Video Sketch, a fine-grained spatio-temporal plan for the final video, specifying background, foreground, and object trajectories, in the form of draft video frames. In the last step, Video-MSG guides a downstream T2V diffusion model with Video Sketch through noise inversion and denoising. Notably, Video-MSG does not need fine-tuning or attention manipulation with additional memory during inference time, making it easier to adopt large T2V models. Video-MSG demonstrates its effectiveness in enhancing text alignment with multiple T2V backbones (VideoCrafter2 and CogVideoX-5B) on popular T2V generation benchmarks (T2VCompBench and VBench). We provide comprehensive ablation studies about noise inversion ratio, different background generators, background object detection, and foreground object segmentation.
Authors: Xinyi Gu, Jiayuan Mao
Abstract: Images not only depict objects but also encapsulate rich interactions between them. However, generating faithful and high-fidelity images involving multiple entities interacting with each other, is a long-standing challenge. While pre-trained text-to-image models are trained on large-scale datasets to follow diverse text instructions, they struggle to generate accurate interactions, likely due to the scarcity of training data for uncommon object interactions. This paper introduces InterActing, an interaction-focused dataset with 1000 fine-grained prompts covering three key scenarios: (1) functional and action-based interactions, (2) compositional spatial relationships, and (3) multi-subject interactions. To address interaction generation challenges, we propose a decomposition-augmented refinement procedure. Our approach, DetailScribe, built on Stable Diffusion 3.5, leverages LLMs to decompose interactions into finer-grained concepts, uses a VLM to critique generated images, and applies targeted interventions within the diffusion process in refinement. Automatic and human evaluations show significantly improved image quality, demonstrating the potential of enhanced inference strategies. Our dataset and code are available at https://concepts-ai.com/p/detailscribe/ to facilitate future exploration of interaction-rich image generation.
Authors: Dayu Yang, Antoine Simoulin, Xin Qian, Xiaoyi Liu, Yuwei Cao, Zhaopu Teng, Grey Yang
Abstract: High-quality code documentation is crucial for software development especially in the era of AI. However, generating it automatically using Large Language Models (LLMs) remains challenging, as existing approaches often produce incomplete, unhelpful, or factually incorrect outputs. We introduce DocAgent, a novel multi-agent collaborative system using topological code processing for incremental context building. Specialized agents (Reader, Searcher, Writer, Verifier, Orchestrator) then collaboratively generate documentation. We also propose a multi-faceted evaluation framework assessing Completeness, Helpfulness, and Truthfulness. Comprehensive experiments show DocAgent significantly outperforms baselines consistently. Our ablation study confirms the vital role of the topological processing order. DocAgent offers a robust approach for reliable code documentation generation in complex and proprietary repositories.
Authors: Yanlin Wang, Kefeng Duan, Dewu Zheng, Ensheng Shi, Fengji Zhang, Yanli Wang, Jiachi Chen, Xilin Liu, Yuchi Ma, Hongyu Zhang, Qianxiang Wang, Zibin Zheng
Abstract: Code intelligence is an emerging domain in software engineering, aiming to improve the effectiveness and efficiency of various code-related tasks. Recent research suggests that incorporating contextual information beyond the basic original task inputs (i.e., source code) can substantially enhance model performance. Such contextual signals may be obtained directly or indirectly from sources such as API documentation or intermediate representations like abstract syntax trees can significantly improve the effectiveness of code intelligence. Despite growing academic interest, there is a lack of systematic analysis of context in code intelligence. To address this gap, we conduct an extensive literature review of 146 relevant studies published between September 2007 and August 2024. Our investigation yields four main contributions. (1) A quantitative analysis of the research landscape, including publication trends, venues, and the explored domains; (2) A novel taxonomy of context types used in code intelligence; (3) A task-oriented analysis investigating context integration strategies across diverse code intelligence tasks; (4) A critical evaluation of evaluation methodologies for context-aware methods. Based on these findings, we identify fundamental challenges in context utilization in current code intelligence systems and propose a research roadmap that outlines key opportunities for future research.
Authors: Jie He, Simon Chi Lok U, V\'ictor Guti\'errez-Basulto, Jeff Z. Pan
Abstract: Unsupervised commonsense reasoning (UCR) is becoming increasingly popular as the construction of commonsense reasoning datasets is expensive, and they are inevitably limited in their scope. A popular approach to UCR is to fine-tune language models with external knowledge (e.g., knowledge graphs), but this usually requires a large number of training examples. In this paper, we propose to transform the downstream multiple choice question answering task into a simpler binary classification task by ranking all candidate answers according to their reasonableness. To this end, for training the model, we convert the knowledge graph triples into reasonable and unreasonable texts. Extensive experimental results show the effectiveness of our approach on various multiple choice question answering benchmarks. Furthermore, compared with existing UCR approaches using KGs, ours is less data hungry. Our code is available at https://github.com/probe2/BUCA.
Authors: Aidan Mannion, Thierry Chevalier, Didier Schwab, Lorraine Geouriot
Abstract: Pre-trained transformer language models (LMs) have in recent years become the dominant paradigm in applied NLP. These models have achieved state-of-the-art performance on tasks such as information extraction, question answering, sentiment analysis, document classification and many others. In the biomedical domain, significant progress has been made in adapting this paradigm to NLP tasks that require the integration of domain-specific knowledge as well as statistical modelling of language. In particular, research in this area has focused on the question of how best to construct LMs that take into account not only the patterns of token distribution in medical text, but also the wealth of structured information contained in terminology resources such as the UMLS. This work contributes a data-centric paradigm for enriching the language representations of biomedical transformer-encoder LMs by extracting text sequences from the UMLS. This allows for graph-based learning objectives to be combined with masked-language pre-training. Preliminary results from experiments in the extension of pre-trained LMs as well as training from scratch show that this framework improves downstream performance on multiple biomedical and clinical Named Entity Recognition (NER) tasks.
Authors: Yash Saxena, Deepa Tilwani, Ali Mohammadi, Edward Raff, Amit Sheth, Srinivasan Parthasarathy, Manas Gaur
Abstract: Large language models (LLMs) present a promising yet challenging frontier for automated source citation in scientific communication. Previous approaches to citation generation have been limited by citation ambiguity and LLM overgeneralization. We introduce REASONS, a novel dataset with sentence-level annotations across 12 scientific domains from arXiv. Our evaluation framework covers two key citation scenarios: indirect queries (matching sentences to paper titles) and direct queries (author attribution), both enhanced with contextual metadata. We conduct extensive experiments with models such as GPT-O1, GPT-4O, GPT-3.5, DeepSeek, and other smaller models like Perplexity AI (7B). While top-tier LLMs achieve high performance in sentence attribution, they struggle with high hallucination rates, a key metric for scientific reliability. Our metadata-augmented approach reduces hallucination rates across all tasks, offering a promising direction for improvement. Retrieval-augmented generation (RAG) with Mistral improves performance in indirect queries, reducing hallucination rates by 42% and maintaining competitive precision with larger models. However, adversarial testing highlights challenges in linking paper titles to abstracts, revealing fundamental limitations in current LLMs. REASONS provides a challenging benchmark for developing reliable and trustworthy LLMs in scientific applications
Authors: Rachneet Sachdeva, Yixiao Song, Mohit Iyyer, Iryna Gurevych
Abstract: Long-form question answering (LFQA) aims to provide thorough and in-depth answers to complex questions, enhancing comprehension. However, such detailed responses are prone to hallucinations and factual inconsistencies, challenging their faithful evaluation. This work introduces HaluQuestQA, the first hallucination dataset with localized error annotations for human-written and model-generated LFQA answers. HaluQuestQA comprises 698 QA pairs with 1.8k span-level error annotations for five different error types by expert annotators, along with preference judgments. Using our collected data, we thoroughly analyze the shortcomings of long-form answers and find that they lack comprehensiveness and provide unhelpful references. We train an automatic feedback model on this dataset that predicts error spans with incomplete information and provides associated explanations. Finally, we propose a prompt-based approach, Error-informed refinement, that uses signals from the learned feedback model to refine generated answers, which we show reduces errors and improves answer quality across multiple models. Furthermore, humans find answers generated by our approach comprehensive and highly prefer them (84%) over the baseline answers.
Authors: Mingning Guo, Mengwei Wu, Yuxiang Shen, Haifeng Li, Chao Tao
Abstract: End-to-end interpretation currently dominates the remote sensing fine-grained ship classification (RS-FGSC) task. However, the inference process remains uninterpretable, leading to criticisms of these models as "black box" systems. To address this issue, we propose a domain knowledge-enhanced Chain-of-Thought (CoT) prompt generation mechanism, which is used to semi-automatically construct a task-specific instruction-following dataset, TITANIC-FGS. By training on TITANIC-FGS, we adapt general-domain vision-language models (VLMs) to the FGSC task, resulting in a model named IFShip. Building upon IFShip, we develop an FGSC visual chatbot that redefines the FGSC problem as a step-by-step reasoning task and conveys the reasoning process in natural language. Experimental results show that IFShip outperforms state-of-the-art FGSC algorithms in both interpretability and classification accuracy. Furthermore, compared to VLMs such as LLaVA and MiniGPT-4, IFShip demonstrates superior performance on the FGSC task. It provides an accurate chain of reasoning when fine-grained ship types are recognizable to the human eye and offers interpretable explanations when they are not.
Authors: Yu Fu, Jie He, Yifan Yang, Qun Liu, Deyi Xiong
Abstract: Meta learning has been widely used to exploit rich-resource source tasks to improve the performance of low-resource target tasks. Unfortunately, most existing meta learning approaches treat different source tasks equally, ignoring the relatedness of source tasks to the target task in knowledge transfer. To mitigate this issue, we propose a reinforcement-based multi-source meta-transfer learning framework (Meta-RTL) for low-resource commonsense reasoning. In this framework, we present a reinforcement-based approach to dynamically estimating source task weights that measure the contribution of the corresponding tasks to the target task in the meta-transfer learning. The differences between the general loss of the meta model and task-specific losses of source-specific temporal meta models on sampled target data are fed into the policy network of the reinforcement learning module as rewards. The policy network is built upon LSTMs that capture long-term dependencies on source task weight estimation across meta learning iterations. We evaluate the proposed Meta-RTL using both BERT and ALBERT as the backbone of the meta model on three commonsense reasoning benchmark datasets. Experimental results demonstrate that Meta-RTL substantially outperforms strong baselines and previous task selection strategies and achieves larger improvements on extremely low-resource settings.
Authors: Haotian Ye, Axel Wisiorek, Antonis Maronikolakis, \"Ozge Ala\c{c}am, Hinrich Sch\"utze
Abstract: Hate speech online remains an understudied issue for marginalized communities, particularly in the Global South, which includes developing societies with increasing internet penetration. In this paper, we aim to provide marginalized communities in societies where the dominant language is low-resource with a privacy-preserving tool to protect themselves from online hate speech by filtering offensive content in their native languages. Our contributions are twofold: 1) we release REACT (REsponsive hate speech datasets Across ConTexts), a collection of high-quality, culture-specific hate speech detection datasets comprising multiple target groups and low-resource languages, curated by experienced data collectors; 2) we propose a few-shot hate speech detection approach based on federated learning (FL), a privacy-preserving method for collaboratively training a central model that exhibits robustness when tackling different target groups and languages. By keeping training local to user devices, we ensure data privacy while leveraging the collective learning benefits of FL. Furthermore, we explore personalized client models tailored to specific target groups and evaluate their performance. Our findings indicate the overall effectiveness of FL across different target groups, and point to personalization as a promising direction.
Authors: Pu Zhao, Xuan Shen, Zhenglun Kong, Yixin Shen, Sung-En Chang, Timothy Rupprecht, Lei Lu, Enfu Nan, Changdi Yang, Yumei He, Weiyan Shi, Xingchen Xu, Yu Huang, Wei Jiang, Wei Wang, Yue Chen, Yong He, Yanzhi Wang
Abstract: Recently, Large Language Models (LLMs) have undergone a significant transformation, marked by a rapid rise in both their popularity and capabilities. Leading this evolution are proprietary LLMs like GPT-4 and GPT-o1, which have captured widespread attention in the AI community due to their remarkable performance and versatility. Simultaneously, open-source LLMs, such as LLaMA, have made great contributions to the ever-increasing popularity of LLMs due to the ease to customize and deploy the models across diverse applications. Although open-source LLMs present unprecedented opportunities for innovation and research, the commercialization of LLMs has raised concerns about transparency, reproducibility, and safety. Many open-source LLMs fail to meet fundamental transparency requirements by withholding essential components like training code and data, which may hinder further innovations on LLMs. To mitigate this issue, we introduce Moxin 7B, a fully open-source LLM developed, adhering to principles of open science, open source, open data, and open access. We release the pre-training code and configurations, training and fine-tuning datasets, and intermediate and final checkpoints, aiming to make continuous commitments to fully open-source LLMs. After pre-training and obtaining the base model, we finetune the Moxin Base model with SOTA post-training framework and instruction data to obtain Moxin Instruct model. To improve the reasoning capability, we further finetune our Instruct model with chain-of-thought data distilled from DeepSeek R1, and then use Group Relative Policy Optimization (GRPO), an efficient and effective reinforcement learning algorithm following DeepSeek R1, to finetune our model, leading to the Moxin Reasoning model. Experiments show that our models achieve superior performance in various evaluations such as zero-shot evaluation, few-shot evaluation, and CoT evaluation.
Authors: Jiankang Wang, Jianjun Xu, Xiaorui Wang, Yuxin Wang, Mengting Xing, Shancheng Fang, Zhineng Chen, Hongtao Xie, Yongdong Zhang
Abstract: Synthesizing high-quality reasoning data for continual training has been proven to be effective in enhancing the performance of Large Language Models (LLMs). However, previous synthetic approaches struggle to easily scale up data and incur high costs in the pursuit of high quality. In this paper, we propose the Graph-based Synthetic Data Pipeline (GSDP), an economical and scalable framework for high-quality reasoning data synthesis. Inspired by knowledge graphs, we extracted knowledge points from seed data and constructed a knowledge point relationships graph to explore their interconnections. By exploring the implicit relationships among knowledge, our method achieves $\times$255 data expansion. Furthermore, GSDP led by open-source models, achieves synthesis quality comparable to GPT-4-0613 while maintaining $\times$100 lower costs. To tackle the most challenging mathematical reasoning task, we present the GSDP-MATH dataset comprising over 1.91 million pairs of math problems and answers. After fine-tuning on GSDP-MATH, GSDP-7B based on Mistral-7B achieves 37.7% accuracy on MATH and 78.4% on GSM8K, demonstrating the effectiveness of our method. The dataset and models will be released at https://github.com/Jayce1kk/GSDP.
Authors: Sieun Hyeon, Kyudan Jung, Jaehee Won, Nam-Joon Kim, Hyun Gon Ryu, Hyuk-Jae Lee, Jaeyoung Do
Abstract: In various academic and professional settings, such as mathematics lectures or research presentations, it is often necessary to convey mathematical expressions orally. However, reading mathematical expressions aloud without accompanying visuals can significantly hinder comprehension, especially for those who are hearing-impaired or rely on subtitles due to language barriers. For instance, when a presenter reads Euler's Formula, current Automatic Speech Recognition (ASR) models often produce a verbose and error-prone textual description (e.g., e to the power of i x equals cosine of x plus i $\textit{side}$ of x), instead of the concise $\LaTeX{}$ format (i.e., $ e^{ix} = \cos(x) + i\sin(x) $), which hampers clear understanding and communication. To address this issue, we introduce MathSpeech, a novel pipeline that integrates ASR models with small Language Models (sLMs) to correct errors in mathematical expressions and accurately convert spoken expressions into structured $\LaTeX{}$ representations. Evaluated on a new dataset derived from lecture recordings, MathSpeech demonstrates $\LaTeX{}$ generation capabilities comparable to leading commercial Large Language Models (LLMs), while leveraging fine-tuned small language models of only 120M parameters. Specifically, in terms of CER, BLEU, and ROUGE scores for $\LaTeX{}$ translation, MathSpeech demonstrated significantly superior capabilities compared to GPT-4o. We observed a decrease in CER from 0.390 to 0.298, and higher ROUGE/BLEU scores compared to GPT-4o.
Authors: Ahmed K. Kadhim, Lei Jiao, Rishad Shafik, Ole-Christoffer Granmo
Abstract: In recent years, text generation tools utilizing Artificial Intelligence (AI) have occasionally been misused across various domains, such as generating student reports or creative writings. This issue prompts plagiarism detection services to enhance their capabilities in identifying AI-generated content. Adversarial attacks are often used to test the robustness of AI-text generated detectors. This work proposes a novel textual adversarial attack on the detection models such as Fast-DetectGPT. The method employs embedding models for data perturbation, aiming at reconstructing the AI generated texts to reduce the likelihood of detection of the true origin of the texts. Specifically, we employ different embedding techniques, including the Tsetlin Machine (TM), an interpretable approach in machine learning for this purpose. By combining synonyms and embedding similarity vectors, we demonstrates the state-of-the-art reduction in detection scores against Fast-DetectGPT. Particularly, in the XSum dataset, the detection score decreased from 0.4431 to 0.2744 AUROC, and in the SQuAD dataset, it dropped from 0.5068 to 0.3532 AUROC.
Authors: Jungsoo Park, Junmo Kang, Gabriel Stanovsky, Alan Ritter
Abstract: The surge of LLM studies makes synthesizing their findings challenging. Analysis of experimental results from literature can uncover important trends across studies, but the time-consuming nature of manual data extraction limits its use. Our study presents a semi-automated approach for literature analysis that accelerates data extraction using LLMs. It automatically identifies relevant arXiv papers, extracts experimental results and related attributes, and organizes them into a structured dataset, LLMEvalDB. We then conduct an automated literature analysis of frontier LLMs, reducing the effort of paper surveying and data extraction by more than 93% compared to manual approaches. We validate LLMEvalDB by showing that it reproduces key findings from a recent manual analysis of Chain-of-Thought (CoT) reasoning and also uncovers new insights that go beyond it, showing, for example, that in-context examples benefit coding and multimodal tasks but offer limited gains in math reasoning tasks compared to zero-shot CoT. Our automatically updatable dataset enables continuous tracking of target models by extracting evaluation studies as new data becomes available. Through LLMEvalDB and empirical analysis, we provide insights into LLMs while facilitating ongoing literature analyses of their behavior.
Authors: So Won Jeong, Veronika Ro\v{c}kov\'a
Abstract: For a long time, the authorship of the Federalist Papers had been a subject of inquiry and debate, not only by linguists and historians but also by statisticians. In what was arguably the first Bayesian case study, Mosteller and Wallace (1963) provided the first statistical evidence for attributing all disputed papers to Madison. Our paper revisits this historical dataset but from a lens of modern language models, both small and large. We review some of the more popular Large Language Model (LLM) tools and examine them from a statistical point of view in the context of text classification. We investigate whether, without any attempt to fine-tune, the general embedding constructs can be useful for stylometry and attribution. We explain differences between various word/phrase embeddings and discuss how to aggregate them in a document. Contrary to our expectations, we exemplify that dimension expansion with word embeddings may not always be beneficial for attribution relative to dimension reduction with topic embeddings. Our experiments demonstrate that default LLM embeddings (even after manual fine-tuning) may not consistently improve authorship attribution accuracy. Instead, Bayesian analysis with topic embeddings trained on ``function words" yields superior out-of-sample classification performance. This suggests that traditional (small) statistical language models, with their interpretability and solid theoretical foundation, can offer significant advantages in authorship attribution tasks. The code used in this analysis is available at github.com/sowonjeong/slm-to-llm
Authors: Navdeep Kaur, Lachlan McPheat, Alessandra Russo, Anthony G Cohn, Pranava Madhyastha
Abstract: In this paper, we examine the use of Conformal Language Modelling (CLM) alongside Answer Set Programming (ASP) to enhance the performance of standard open-weight LLMs on complex multi-step reasoning tasks. Using the StepGame dataset, which requires spatial reasoning, we apply CLM to generate sets of ASP programs from an LLM, providing statistical guarantees on the correctness of the outputs. Experimental results show that CLM significantly outperforms baseline models that use standard sampling methods, achieving substantial accuracy improvements across different levels of reasoning complexity. Additionally, the LLM-as-Judge metric enhances CLM's performance, especially in assessing structurally and logically correct ASP outputs. However, calibrating CLM with diverse calibration sets did not improve generalizability for tasks requiring much longer reasoning steps, indicating limitations in handling more complex tasks.
Authors: Hongchao Fang, Yixin Liu, Jiangshu Du, Can Qin, Ran Xu, Feng Liu, Lichao Sun, Dongwon Lee, Lifu Huang, Wenpeng Yin
Abstract: AI-generated content is becoming increasingly prevalent in the real world, leading to serious ethical and societal concerns. For instance, adversaries might exploit large multimodal models (LMMs) to create images that violate ethical or legal standards, while paper reviewers may misuse large language models (LLMs) to generate reviews without genuine intellectual effort. While prior work has explored detecting AI-generated images and texts, and occasionally tracing their source models, there is a lack of a systematic and fine-grained comparative study. Important dimensions--such as AI-generated images vs. text, fully vs. partially AI-generated images, and general vs. malicious use cases--remain underexplored. Furthermore, whether AI systems like GPT-4o can explain why certain forged content is attributed to specific generative models is still an open question, with no existing benchmark addressing this. To fill this gap, we introduce AI-FAKER, a comprehensive multimodal dataset with over 280,000 samples spanning multiple LLMs and LMMs, covering both general and malicious use cases for AI-generated images and texts. Our experiments reveal two key findings: (i) AI authorship detection depends not only on the generated output but also on the model's original training intent; and (ii) GPT-4o provides highly consistent but less specific explanations when analyzing content produced by OpenAI's own models, such as DALL-E and GPT-4o itself.
Authors: Aravind Krishnan, Siva Reddy, Marius Mosbach
Abstract: Machine unlearning is concerned with the task of removing knowledge learned from particular data points from a trained model. In the context of large language models (LLMs), unlearning has recently received increased attention, particularly for removing knowledge about named entities from models for privacy purposes. While various approaches have been proposed to address the unlearning problem, most existing approaches treat all data points to be unlearned equally, i.e., unlearning that Montreal is a city in Canada is treated exactly the same as unlearning the phone number of the first author of this paper. In this work, we show that this all data is equal assumption does not hold for LLM unlearning. We study how the success of unlearning depends on the frequency of the knowledge we want to unlearn in the pre-training data of a model and find that frequency strongly affects unlearning, i.e., more frequent knowledge is harder to unlearn. Additionally, we uncover a misalignment between probability and generation-based evaluations of unlearning and show that this problem worsens as models become larger. Overall, our experiments highlight the need for better evaluation practices and novel methods for LLM unlearning that take the training data of models into account.
Authors: Jennifer D'Souza, Sameer Sadruddin, Holger Israel, Mathias Begoin, Diana Slawig
Abstract: We present SemEval-2025 Task 5: LLMs4Subjects, a shared task on automated subject tagging for scientific and technical records in English and German using the GND taxonomy. Participants developed LLM-based systems to recommend top-k subjects, evaluated through quantitative metrics (precision, recall, F1-score) and qualitative assessments by subject specialists. Results highlight the effectiveness of LLM ensembles, synthetic data generation, and multilingual processing, offering insights into applying LLMs for digital library classification.
Authors: Patrick Fernandes, Sweta Agrawal, Emmanouil Zaranis, Andr\'e F. T. Martins, Graham Neubig
Abstract: Despite the steady progress in machine translation evaluation, existing automatic metrics struggle to capture how well meaning is preserved beyond sentence boundaries. We posit that reliance on a single intrinsic quality score, trained to mimic human judgments, might be insufficient for evaluating translations of long, complex passages, and a more ``pragmatic'' approach that assesses how accurately key information is conveyed by a translation in context is needed. We introduce TREQA (Translation Evaluation via Question-Answering), a framework that extrinsically evaluates translation quality by assessing how accurately candidate translations answer reading comprehension questions that target key information in the original source or reference texts. In challenging domains that require long-range understanding, such as literary texts, we show that TREQA is competitive with and, in some cases, outperforms state-of-the-art neural and LLM-based metrics in ranking alternative paragraph-level translations, despite never being explicitly optimized to correlate with human judgments. Furthermore, the generated questions and answers offer interpretability: empirical analysis shows that they effectively target translation errors identified by experts in evaluated datasets. Our code is available at https://github.com/deep-spin/treqa
Authors: Yichun Yin, Wenyong Huang, Kaikai Song, Yehui Tang, Xueyu Wu, Wei Guo, Peng Guo, Yaoyuan Wang, Xiaojun Meng, Yasheng Wang, Dong Li, Can Chen, Dandan Tu, Yin Li, Fisher Yu, Ruiming Tang, Yunhe Wang, Baojun Wang, Bin Wang, Bo Wang, Boxiao Liu, Changzheng Zhang, Duyu Tang, Fei Mi, Hui Jin, Jiansheng Wei, Jiarui Qin, Jinpeng Li, Jun Zhao, Liqun Deng, Lin Li, Minghui Xu, Naifu Zhang, Nianzu Zheng, Qiang Li, Rongju Ruan, Shengjun Cheng, Tianyu Guo, Wei He, Wei Li, Weiwen Liu, Wulong Liu, Xinyi Dai, Yonghan Dong, Yu Pan, Yue Li, Yufei Wang, Yujun Li, Yunsheng Ni, Zhe Liu, Zhenhe Zhang, Zhicheng Liu
Abstract: We present Pangu Ultra, a Large Language Model (LLM) with 135 billion parameters and dense Transformer modules trained on Ascend Neural Processing Units (NPUs). Although the field of LLM has been witnessing unprecedented advances in pushing the scale and capability of LLM in recent years, training such a large-scale model still involves significant optimization and system challenges. To stabilize the training process, we propose depth-scaled sandwich normalization, which effectively eliminates loss spikes during the training process of deep models. We pre-train our model on 13.2 trillion diverse and high-quality tokens and further enhance its reasoning capabilities during post-training. To perform such large-scale training efficiently, we utilize 8,192 Ascend NPUs with a series of system optimizations. Evaluations on multiple diverse benchmarks indicate that Pangu Ultra significantly advances the state-of-the-art capabilities of dense LLMs such as Llama 405B and Mistral Large 2, and even achieves competitive results with DeepSeek-R1, whose sparse model structure contains much more parameters. Our exploration demonstrates that Ascend NPUs are capable of efficiently and effectively training dense models with more than 100 billion parameters. Our model and system will be available for our commercial customers.
Authors: Haoxuan You, Zhecan Wang, Rui Sun, Long Chen, Gengyu Wang, Hammad A. Ayyubi, Kai-Wei Chang, Shih-Fu Chang
Abstract: The field of vision-and-language (VL) understanding has made unprecedented progress with end-to-end large pre-trained VL models (VLMs). However, they still fall short in zero-shot reasoning tasks that require multi-step inferencing. To achieve this goal, previous works resort to a divide-and-conquer pipeline. In this paper, we argue that previous efforts have several inherent shortcomings: 1) They rely on domain-specific sub-question decomposing models. 2) They force models to predict the final answer even if the sub-questions or sub-answers provide insufficient information. We address these limitations via IdealGPT, a framework that iteratively decomposes VL reasoning using large language models (LLMs). Specifically, IdealGPT utilizes an LLM to generate sub-questions, a VLM to provide corresponding sub-answers, and another LLM to reason to achieve the final answer. These three modules perform the divide-and-conquer procedure iteratively until the model is confident about the final answer to the main question. We evaluate IdealGPT on multiple challenging VL reasoning tasks under a zero-shot setting. In particular, our IdealGPT outperforms the best existing GPT-4-like models by an absolute 10% on VCR and 15% on SNLI-VE. Code is available at https://github.com/Hxyou/IdealGPT
Authors: Ziteng Sun, Uri Mendlovic, Yaniv Leviathan, Asaf Aharoni, Jae Hun Ro, Ahmad Beirami, Ananda Theertha Suresh
Abstract: Speculative decoding is an effective method for lossless acceleration of large language models during inference. It uses a fast model to draft a block of tokens which are then verified in parallel by the target model, and provides a guarantee that the output is distributed identically to a sample from the target model. In prior works, draft verification is performed independently token-by-token. Surprisingly, we show that this approach is not optimal. We propose Block Verification, a simple draft verification algorithm that verifies the entire block jointly and provides additional wall-clock speedup. We prove that the proposed mechanism is optimal in the expected number of tokens produced each iteration and specifically is never worse than the standard token-level verification. Empirically, block verification provides modest but consistent wall-clock speedups over the standard token verification algorithm of 5%-8% in a range of tasks and datasets. Given that block verification does not increase code complexity, maintains the strong lossless guarantee of the standard speculative decoding verification algorithm, cannot deteriorate performance, and, in fact, consistently improves it, it can be used as a good default in speculative decoding implementations.
Authors: Liqiang Jing, Zhehui Huang, Xiaoyang Wang, Wenlin Yao, Wenhao Yu, Kaixin Ma, Hongming Zhang, Xinya Du, Dong Yu
Abstract: Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) have demonstrated impressive language/vision reasoning abilities, igniting the recent trend of building agents for targeted applications such as shopping assistants or AI software engineers. Recently, many data science benchmarks have been proposed to investigate their performance in the data science domain. However, existing data science benchmarks still fall short when compared to real-world data science applications due to their simplified settings. To bridge this gap, we introduce DSBench, a comprehensive benchmark designed to evaluate data science agents with realistic tasks. This benchmark includes 466 data analysis tasks and 74 data modeling tasks, sourced from Eloquence and Kaggle competitions. DSBench offers a realistic setting by encompassing long contexts, multimodal task backgrounds, reasoning with large data files and multi-table structures, and performing end-to-end data modeling tasks. Our evaluation of state-of-the-art LLMs, LVLMs, and agents shows that they struggle with most tasks, with the best agent solving only 34.12% of data analysis tasks and achieving a 34.74% Relative Performance Gap (RPG). These findings underscore the need for further advancements in developing more practical, intelligent, and autonomous data science agents.
Authors: Zhili Cheng, Yuge Tu, Ran Li, Shiqi Dai, Jinyi Hu, Shengding Hu, Jiahao Li, Yang Shi, Tianyu Yu, Weize Chen, Lei Shi, Maosong Sun
Abstract: Multimodal Large Language Models (MLLMs) have shown significant advancements, providing a promising future for embodied agents. Existing benchmarks for evaluating MLLMs primarily utilize static images or videos, limiting assessments to non-interactive scenarios. Meanwhile, existing embodied AI benchmarks are task-specific and not diverse enough, which do not adequately evaluate the embodied capabilities of MLLMs. To address this, we propose EmbodiedEval, a comprehensive and interactive evaluation benchmark for MLLMs with embodied tasks. EmbodiedEval features 328 distinct tasks within 125 varied 3D scenes, each of which is rigorously selected and annotated. It covers a broad spectrum of existing embodied AI tasks with significantly enhanced diversity, all within a unified simulation and evaluation framework tailored for MLLMs. The tasks are organized into five categories: navigation, object interaction, social interaction, attribute question answering, and spatial question answering to assess different capabilities of the agents. We evaluated the state-of-the-art MLLMs on EmbodiedEval and found that they have a significant shortfall compared to human level on embodied tasks. Our analysis demonstrates the limitations of existing MLLMs in embodied capabilities, providing insights for their future development. We open-source all evaluation data and simulation framework at https://github.com/thunlp/EmbodiedEval.
Authors: Xuefeng Liu, Songhao Jiang, Ian Foster, Jinbo Xu, Rick Stevens
Abstract: Drug optimization has become increasingly crucial in light of fast-mutating virus strains and drug-resistant cancer cells. Nevertheless, it remains challenging as it necessitates retaining the beneficial properties of the original drug while simultaneously enhancing desired attributes beyond its scope. In this work, we aim to tackle this challenge by introducing ScaffoldGPT, a novel Generative Pretrained Transformer (GPT) designed for drug optimization based on molecular scaffolds. Our work comprises three key components: (1) A three-stage drug optimization approach that integrates pretraining, finetuning, and decoding optimization. (2) A uniquely designed two-phase incremental training approach for pre-training the drug optimization GPT on molecule scaffold with enhanced performance. (3) A token-level decoding optimization strategy, TOP-N, that enabling controlled, reward-guided generation using pretrained/finetuned GPT. We demonstrate via a comprehensive evaluation on COVID and cancer benchmarks that ScaffoldGPT outperforms the competing baselines in drug optimization benchmarks, while excelling in preserving original functional scaffold and enhancing desired properties.
Authors: Simeng Sun, Cheng-Ping Hsieh, Faisal Ladhak, Erik Arakelyan, Santiago Akle Serano, Boris Ginsburg
Abstract: Complex reasoning tasks often rely on the ability to consistently and accurately apply simple rules across incremental steps, a foundational capability which we term "level-0" reasoning. To systematically evaluate this capability, we introduce L0-Bench, a language model benchmark for testing procedural correctness -- the ability to generate correct reasoning processes, complementing existing benchmarks that primarily focus on outcome correctness. Given synthetic Python functions with simple operations, L0-Bench grades models on their ability to generate step-by-step, error-free execution traces. The synthetic nature of L0-Bench enables systematic and scalable generation of test programs along various axes (e.g., number of trace steps). We evaluate a diverse array of recent closed-source and open-weight models on a baseline test set. All models exhibit degradation as the number of target trace steps increases, while larger models and reasoning-enhanced models better maintain correctness over multiple steps. Additionally, we use L0-Bench to explore test-time scaling along three dimensions: input context length, number of solutions for majority voting, and inference steps. Our results suggest substantial room to improve "level-0" reasoning and potential directions to build more reliable reasoning systems.
Authors: Chenrui Fan, Ming Li, Lichao Sun, Tianyi Zhou
Abstract: We find that the response length of reasoning LLMs, whether trained by reinforcement learning or supervised learning, drastically increases for ill-posed questions with missing premises (MiP), ending up with redundant and ineffective thinking. This newly introduced scenario exacerbates the general overthinking issue to a large extent, which we name as the MiP-Overthinking. Such failures are against the ``test-time scaling law'' but have been widely observed on multiple datasets we curated with MiP, indicating the harm of cheap overthinking and a lack of critical thinking. Surprisingly, LLMs not specifically trained for reasoning exhibit much better performance on the MiP scenario, producing much shorter responses that quickly identify ill-posed queries. This implies a critical flaw of the current training recipe for reasoning LLMs, which does not encourage efficient thinking adequately, leading to the abuse of thinking patterns. To further investigate the reasons behind such failures, we conduct fine-grained analyses of the reasoning length, overthinking patterns, and location of critical thinking on different types of LLMs. Moreover, our extended ablation study reveals that the overthinking is contagious through the distillation of reasoning models' responses. These results improve the understanding of overthinking and shed novel insights into mitigating the problem.
Authors: Ling Team, Caizhi Tang, Chilin Fu, Chunwei Wu, Jia Guo, Jianwen Wang, Jingyu Hu, Liang Jiang, Meng Li, Peng Jiao, Pingping Liu, Shaomian Zheng, Shiwei Liang, Shuaicheng Li, Yalin Zhang, Yingting Wu, Yongkang Liu, Zhenyu Huang
Abstract: This technical report presents Ring-Lite-Distill, a lightweight reasoning model derived from our open-source Mixture-of-Experts (MoE) Large Language Models (LLMs) Ling-Lite. This study demonstrates that through meticulous high-quality data curation and ingenious training paradigms, the compact MoE model Ling-Lite can be further trained to achieve exceptional reasoning capabilities, while maintaining its parameter-efficient architecture with only 2.75 billion activated parameters, establishing an efficient lightweight reasoning architecture. In particular, in constructing this model, we have not merely focused on enhancing advanced reasoning capabilities, exemplified by high-difficulty mathematical problem solving, but rather aimed to develop a reasoning model with more comprehensive competency coverage. Our approach ensures coverage across reasoning tasks of varying difficulty levels while preserving generic capabilities, such as instruction following, tool use, and knowledge retention. We show that, Ring-Lite-Distill's reasoning ability reaches a level comparable to DeepSeek-R1-Distill-Qwen-7B, while its general capabilities significantly surpass those of DeepSeek-R1-Distill-Qwen-7B. The models are accessible at https://huggingface.co/inclusionAI