Authors: Seanghort Born, Madeth May, Claudine Piau-Toffolon, S\'ebastien Iksal
Abstract: Homophones present a significant challenge to authors in any languages due to their similarities of pronunciations but different meanings and spellings. This issue is particularly pronounced in the Khmer language, rich in homophones due to its complex structure and extensive character set. This research aims to address the difficulties faced by Khmer authors when using homophones in their writing and proposes potential solutions based on an extensive literature review and survey analysis. A survey of 108 Khmer native speakers, including students, employees, and professionals, revealed that many frequently encounter challenges with homophones in their writing, often struggling to choose the correct word based on context. The survey also highlighted the absence of effective tools to address homophone errors in Khmer, which complicates the writing process. Additionally, a review of existing studies on spelling correction in other languages, such as English, Azerbaijani, and Bangla, identified a lack of research focused specifically on homophones, particularly in the Khmer language. In summary, this research highlights the necessity for a specialized tool to address Khmer homophone errors. By bridging current gaps in research and available resources, such a tool would enhance the confidence and accuracy of Khmer authors in their writing, thereby contributing to the enrichment and preservation of the language. Continued efforts in this domain are essential for ensuring that Khmer can leverage advancements in technology and linguistics effectively.
Authors: Eva Portelance, Masoud Jasbi
Abstract: Theoretical linguistics seeks to explain what human language is, and why. Linguists and cognitive scientists have proposed different theoretical models of what language is, as well as cognitive factors that shape it, and allow humans to 'produce', 'understand', and 'acquire' natural languages. However, humans may no longer be the only ones learning to 'generate', 'parse', and 'learn' natural language: artificial intelligence (AI) models such as large language models are proving to have impressive linguistic capabilities. Many are thus questioning what role, if any, such models should play in helping theoretical linguistics reach its ultimate research goals? In this paper, we propose to answer this question, by reiterating the tenets of generative linguistics, a leading school of thought in the field, and by considering how AI models as theories of language relate to each of these important concepts. Specifically, we consider three foundational principles, finding roots in the early works of Noam Chomsky: (1) levels of theoretical adequacy; (2) procedures for linguistic theory development; (3) language learnability and Universal Grammar. In our discussions of each principle, we give special attention to two types of AI models: neural language models and neural grammar induction models. We will argue that such models, in particular neural grammar induction models, do have a role to play, but that this role is largely modulated by the stance one takes regarding each of these three guiding principles.
Authors: Jia He, Mukund Rungta, David Koleczek, Arshdeep Sekhon, Franklin X Wang, Sadid Hasan
Abstract: In the realm of Large Language Models (LLMs), prompt optimization is crucial for model performance. Although previous research has explored aspects like rephrasing prompt contexts, using various prompting techniques (like in-context learning and chain-of-thought), and ordering few-shot examples, our understanding of LLM sensitivity to prompt templates remains limited. Therefore, this paper examines the impact of different prompt templates on LLM performance. We formatted the same contexts into various human-readable templates, including plain text, Markdown, JSON, and YAML, and evaluated their impact across tasks like natural language reasoning, code generation, and translation using OpenAI's GPT models. Experiments show that GPT-3.5-turbo's performance varies by up to 40\% in a code translation task depending on the prompt template, while larger models like GPT-4 are more robust to these variations. Our analysis highlights the need to reconsider the use of fixed prompt templates, as different formats can significantly affect model performance.
Authors: Jianhong Tu, Zhuohao Ni, Nicholas Crispino, Zihao Yu, Michael Bendersky, Beliz Gunel, Ruoxi Jia, Xin Liu, Lingjuan Lyu, Dawn Song, Chenguang Wang
Abstract: We present a novel instruction tuning recipe to improve the zero-shot task generalization of multimodal large language models. In contrast to existing instruction tuning mechanisms that heavily rely on visual instructions, our approach focuses on language-based instruction tuning, offering a distinct and more training efficient path for multimodal instruction tuning. We evaluate the performance of the proposed approach on 9 unseen datasets across both language and vision modalities. Our results show that our language-only instruction tuning is able to significantly improve the performance of two pretrained multimodal models based on Llama 2 and Vicuna on those unseen datasets. Interestingly, the language instruction following ability also helps unlock the models to follow vision instructions without explicit training. Compared to the state of the art multimodal instruction tuning approaches that are mainly based on visual instructions, our language-based method not only achieves superior performance but also significantly enhances training efficiency. For instance, the language-only instruction tuning produces competitive average performance across the evaluated datasets (with even better performance on language datasets) with significant training efficiency improvements (on average 4x), thanks to the striking reduction in the need for vision data. With a small number of visual instructions, this emerging language instruction following ability transfers well to the unseen vision datasets, outperforming the state of the art with greater training efficiency.
Authors: Wenxuan Wang, Wenxiang Jiao, Jen-tse Huang, Zhaopeng Tu, Michael R. Lyu
Abstract: In this study, we revisit the commonly-cited off-target issue in multilingual neural machine translation (MNMT). By carefully designing experiments on different MNMT scenarios and models, we attribute the off-target issue to the overfitting of the shortcuts of (non-centric, centric) language mappings. Specifically, the learned shortcuts biases MNMT to mistakenly translate non-centric languages into the centric language instead of the expected non-centric language for zero-shot translation. Analyses on learning dynamics show that the shortcut learning generally occurs in the later stage of model training, and multilingual pretraining accelerates and aggravates the shortcut learning. Based on these observations, we propose a simple and effective training strategy to eliminate the shortcuts in MNMT models by leveraging the forgetting nature of model training. The only difference from the standard training is that we remove the training instances that may induce the shortcut learning in the later stage of model training. Without introducing any additional data and computational costs, our approach can consistently and significantly improve the zero-shot translation performance by alleviating the shortcut learning for different MNMT models and benchmarks.
Authors: Caspar Oesterheld, Emery Cooper, Miles Kodama, Linh Chi Nguyen, Ethan Perez
Abstract: We introduce a dataset of natural-language questions in the decision theory of so-called Newcomb-like problems. Newcomb-like problems include, for instance, decision problems in which an agent interacts with a similar other agent, and thus has to reason about the fact that the other agent will likely reason in similar ways. Evaluating LLM reasoning about Newcomb-like problems is important because interactions between foundation-model-based agents will often be Newcomb-like. Some ways of reasoning about Newcomb-like problems may allow for greater cooperation between models. Our dataset contains both capabilities questions (i.e., questions with a unique, uncontroversially correct answer) and attitude questions (i.e., questions about which decision theorists would disagree). We use our dataset for an investigation of decision-theoretical capabilities and expressed attitudes and their interplay in existing models (different models by OpenAI, Anthropic, Meta, GDM, Reka, etc.), as well as models under simple prompt-based interventions. We find, among other things, that attitudes vary significantly between existing models; that high capabilities are associated with attitudes more favorable toward so-called evidential decision theory; and that attitudes are consistent across different types of questions.
Authors: Xiaowei Xu, Bi T. Foua, Xingqiao Wang, Vivek Gunasekaran, John R. Talburt
Abstract: In this paper, the authors propose TriBERTa, a supervised entity resolution system that utilizes a pre-trained large language model and a triplet loss function to learn representations for entity matching. The system consists of two steps: first, name entity records are fed into a Sentence Bidirectional Encoder Representations from Transformers (SBERT) model to generate vector representations, which are then fine-tuned using contrastive learning based on a triplet loss function. Fine-tuned representations are used as input for entity matching tasks, and the results show that the proposed approach outperforms state-of-the-art representations, including SBERT without fine-tuning and conventional Term Frequency-Inverse Document Frequency (TF-IDF), by a margin of 3 - 19%. Additionally, the representations generated by TriBERTa demonstrated increased robustness, maintaining consistently higher performance across a range of datasets. The authors also discussed the importance of entity resolution in today's data-driven landscape and the challenges that arise when identifying and reconciling duplicate data across different sources. They also described the ER process, which involves several crucial steps, including blocking, entity matching, and clustering.
Authors: Sajib Kumar Saha Joy, Arman Hassan Mahy, Meherin Sultana, Azizah Mamun Abha, MD Piyal Ahmmed, Yue Dong, G M Shahariar
Abstract: In this study, we investigate gender bias in Bangla pretrained language models, a largely under explored area in low-resource languages. To assess this bias, we applied gender-name swapping techniques to existing datasets, creating four manually annotated, task-specific datasets for sentiment analysis, toxicity detection, hate speech detection, and sarcasm detection. By altering names and gender-specific terms, we ensured these datasets were suitable for detecting and mitigating gender bias. We then proposed a joint loss optimization technique to mitigate gender bias across task-specific pretrained models. Our approach was evaluated against existing bias mitigation methods, with results showing that our technique not only effectively reduces bias but also maintains competitive accuracy compared to other baseline approaches. To promote further research, we have made both our implementation and datasets publicly available https://github.com/sajib-kumar/Gender-Bias-Mitigation-From-Bangla-PLM
URLs: https://github.com/sajib-kumar/Gender-Bias-Mitigation-From-Bangla-PLM
Authors: Yuxuan Hu, Ke Wang, Jing Zhang, Cuiping Li, Hong Chen
Abstract: Large Language Models (LLMs) have revolutionized natural language processing by unifying tasks into text generation, yet their large parameter sizes and autoregressive nature limit inference speed. SAM-Decoding addresses this by introducing a novel retrieval-based speculative decoding method that uses a suffix automaton for efficient and accurate draft generation. Unlike n-gram matching used by the existing method, SAM-Decoding finds the longest suffix match in generating text and text corpuss, achieving an average time complexity of $O(1)$ per generation step. SAM-Decoding constructs static and dynamic suffix automatons for the text corpus and input prompts, respectively, enabling fast and precise draft generation. Meanwhile, it is designed as an approach that can be combined with existing methods, allowing SAM-Decoding to adaptively select a draft generation strategy based on the matching length, thus increasing the inference speed of the LLM. When combined with Token Recycling, evaluations show SAM-Decoding outperforms existing model-free methods, achieving a speedup of $2.27\times$ over autoregressive decoding on Spec-Bench. When combined with EAGLE2, it reaches a speedup of $2.49\times$, surpassing all current approaches. Our code is available at https://github.com/hyx1999/SAM-Decoding.
Authors: Juan A. Rodriguez, Nicholas Botzer, David Vazquez, Christopher Pal, Marco Pedersoli, Issam Laradji
Abstract: In today's digitally driven world, dialogue systems play a pivotal role in enhancing user interactions, from customer service to virtual assistants. In these dialogues, it is important to identify user's goals automatically to resolve their needs promptly. This has necessitated the integration of models that perform Intent Detection. However, users' intents are diverse and dynamic, making it challenging to maintain a fixed set of predefined intents. As a result, a more practical approach is to develop a model capable of identifying new intents as they emerge. We address the challenge of Intent Discovery, an area that has drawn significant attention in recent research efforts. Existing methods need to train on a substantial amount of data for correctly identifying new intents, demanding significant human effort. To overcome this, we introduce IntentGPT, a novel training-free method that effectively prompts Large Language Models (LLMs) such as GPT-4 to discover new intents with minimal labeled data. IntentGPT comprises an \textit{In-Context Prompt Generator}, which generates informative prompts for In-Context Learning, an \textit{Intent Predictor} for classifying and discovering user intents from utterances, and a \textit{Semantic Few-Shot Sampler} that selects relevant few-shot examples and a set of known intents to be injected into the prompt. Our experiments show that IntentGPT outperforms previous methods that require extensive domain-specific data and fine-tuning, in popular benchmarks, including CLINC and BANKING, among others.
Authors: Yixiang Chen, Xinyu Zhang, Jinran Wang, Xurong Xie, Nan Yan, Hui Chen, Lan Wang
Abstract: The Structured Dialogue System, referred to as SuDoSys, is an innovative Large Language Model (LLM)-based chatbot designed to provide psychological counseling. SuDoSys leverages the World Health Organization (WHO)'s Problem Management Plus (PM+) guidelines to deliver stage-aware multi-turn dialogues. Existing methods for employing an LLM in multi-turn psychological counseling typically involve direct fine-tuning using generated dialogues, often neglecting the dynamic stage shifts of counseling sessions. Unlike previous approaches, SuDoSys considers the different stages of counseling and stores essential information throughout the counseling process, ensuring coherent and directed conversations. The system employs an LLM, a stage-aware instruction generator, a response unpacker, a topic database, and a stage controller to maintain dialogue flow. In addition, we propose a novel technique that simulates counseling clients to interact with the evaluated system and evaluate its performance automatically. When assessed using both objective and subjective evaluations, SuDoSys demonstrates its effectiveness in generating logically coherent responses. The system's code and program scripts for evaluation are open-sourced.
Authors: Anton Alekseev, Gulnara Kabaeva
Abstract: One of the key tasks in modern applied computational linguistics is constructing word vector representations (word embeddings), which are widely used to address natural language processing tasks such as sentiment analysis, information extraction, and more. To choose an appropriate method for generating these word embeddings, quality assessment techniques are often necessary. A standard approach involves calculating distances between vectors for words with expert-assessed 'similarity'. This work introduces the first 'silver standard' dataset for such tasks in the Kyrgyz language, alongside training corresponding models and validating the dataset's suitability through quality evaluation metrics.
Authors: Omar W. Abdalla, Aditya Joshi, Rahat Masood, Salil S. Kanhere
Abstract: Satirical news is real news combined with a humorous comment or exaggerated content, and it often mimics the format and style of real news. However, satirical news is often misunderstood as misinformation, especially by individuals from different cultural and social backgrounds. This research addresses the challenge of distinguishing satire from truthful news by leveraging multilingual satire detection methods in English and Arabic. We explore both zero-shot and chain-of-thought (CoT) prompting using two language models, Jais-chat(13B) and LLaMA-2-chat(7B). Our results show that CoT prompting offers a significant advantage for the Jais-chat model over the LLaMA-2-chat model. Specifically, Jais-chat achieved the best performance, with an F1-score of 80\% in English when using CoT prompting. These results highlight the importance of structured reasoning in CoT, which enhances contextual understanding and is vital for complex tasks like satire detection.
Authors: Tiantian Feng, Anfeng Xu, Rimita Lahiri, Helen Tager-Flusberg, So Hyun Kim, Somer Bishop, Catherine Lord, Shrikanth Narayanan
Abstract: Large Language Models (LLMs) have shown significant potential in understanding human communication and interaction. However, their performance in the domain of child-inclusive interactions, including in clinical settings, remains less explored. In this work, we evaluate generic LLMs' ability to analyze child-adult dyadic interactions in a clinically relevant context involving children with ASD. Specifically, we explore LLMs in performing four tasks: classifying child-adult utterances, predicting engaged activities, recognizing language skills and understanding traits that are clinically relevant. Our evaluation shows that generic LLMs are highly capable of analyzing long and complex conversations in clinical observation sessions, often surpassing the performance of non-expert human evaluators. The results show their potential to segment interactions of interest, assist in language skills evaluation, identify engaged activities, and offer clinical-relevant context for assessments.
Authors: Prasoon Bajpai, Sarah Masud, Tanmoy Chakraborty
Abstract: Large Language Models (LLMs) have demonstrated strong performance as knowledge repositories, enabling models to understand user queries and generate accurate and context-aware responses. Extensive evaluation setups have corroborated the positive correlation between the retrieval capability of LLMs and the frequency of entities in their pretraining corpus. We take the investigation further by conducting a comprehensive analysis of the internal reasoning and retrieval mechanisms of LLMs. Our work focuses on three critical dimensions - the impact of entity popularity, the models' sensitivity to lexical variations in query formulation, and the progression of hidden state representations across LLM layers. Our preliminary findings reveal that popular questions facilitate early convergence of internal states toward the correct answer. However, as the popularity of a query increases, retrieved attributes across lexical variations become increasingly dissimilar and less accurate. Interestingly, we find that LLMs struggle to disentangle facts, grounded in distinct relations, from their parametric memory when dealing with highly popular subjects. Through a case study, we explore these latent strains within LLMs when processing highly popular queries, a phenomenon we term information anxiety. The emergence of information anxiety in LLMs underscores the adversarial injection in the form of linguistic variations and calls for a more holistic evaluation of frequently occurring entities.
Authors: Sari Masri, Huthaifa I. Ashqar, Mohammed Elhenawy
Abstract: This study introduces a novel approach for traffic control systems by using Large Language Models (LLMs) as traffic controllers. The study utilizes their logical reasoning, scene understanding, and decision-making capabilities to optimize throughput and provide feedback based on traffic conditions in real-time. LLMs centralize traditionally disconnected traffic control processes and can integrate traffic data from diverse sources to provide context-aware decisions. LLMs can also deliver tailored outputs using various means such as wireless signals and visuals to drivers, infrastructures, and autonomous vehicles. To evaluate LLMs ability as traffic controllers, this study proposed a four-stage methodology. The methodology includes data creation and environment initialization, prompt engineering, conflict identification, and fine-tuning. We simulated multi-lane four-leg intersection scenarios and generates detailed datasets to enable conflict detection using LLMs and Python simulation as a ground truth. We used chain-of-thought prompts to lead LLMs in understanding the context, detecting conflicts, resolving them using traffic rules, and delivering context-sensitive traffic management solutions. We evaluated the prformance GPT-mini, Gemini, and Llama as traffic controllers. Results showed that the fine-tuned GPT-mini achieved 83% accuracy and an F1-score of 0.84. GPT-mini model exhibited a promising performance in generating actionable traffic management insights, with high ROUGE-L scores across conflict identification of 0.95, decision-making of 0.91, priority assignment of 0.94, and waiting time optimization of 0.92. We demonstrated that LLMs can offer precise recommendations to drivers in real-time including yielding, slowing, or stopping based on vehicle dynamics.
Authors: Jawad Ibn Ahad, Rafeed Mohammad Sultan, Abraham Kaikobad, Fuad Rahman, Mohammad Ruhul Amin, Nabeel Mohammed, Shafin Rahman
Abstract: This study investigates the automation of meta-analysis in scientific documents using large language models (LLMs). Meta-analysis is a robust statistical method that synthesizes the findings of multiple studies support articles to provide a comprehensive understanding. We know that a meta-article provides a structured analysis of several articles. However, conducting meta-analysis by hand is labor-intensive, time-consuming, and susceptible to human error, highlighting the need for automated pipelines to streamline the process. Our research introduces a novel approach that fine-tunes the LLM on extensive scientific datasets to address challenges in big data handling and structured data extraction. We automate and optimize the meta-analysis process by integrating Retrieval Augmented Generation (RAG). Tailored through prompt engineering and a new loss metric, Inverse Cosine Distance (ICD), designed for fine-tuning on large contextual datasets, LLMs efficiently generate structured meta-analysis content. Human evaluation then assesses relevance and provides information on model performance in key metrics. This research demonstrates that fine-tuned models outperform non-fine-tuned models, with fine-tuned LLMs generating 87.6% relevant meta-analysis abstracts. The relevance of the context, based on human evaluation, shows a reduction in irrelevancy from 4.56% to 1.9%. These experiments were conducted in a low-resource environment, highlighting the study's contribution to enhancing the efficiency and reliability of meta-analysis automation.
Authors: Md. Nazmus Sadat Samin, Jawad Ibn Ahad, Tanjila Ahmed Medha, Fuad Rahman, Mohammad Ruhul Amin, Nabeel Mohammed, Shafin Rahman
Abstract: This study focuses on recognizing Bangladeshi dialects and converting diverse Bengali accents into standardized formal Bengali speech. Dialects, often referred to as regional languages, are distinctive variations of a language spoken in a particular location and are identified by their phonetics, pronunciations, and lexicon. Subtle changes in pronunciation and intonation are also influenced by geographic location, educational attainment, and socioeconomic status. Dialect standardization is needed to ensure effective communication, educational consistency, access to technology, economic opportunities, and the preservation of linguistic resources while respecting cultural diversity. Being the fifth most spoken language with around 55 distinct dialects spoken by 160 million people, addressing Bangla dialects is crucial for developing inclusive communication tools. However, limited research exists due to a lack of comprehensive datasets and the challenges of handling diverse dialects. With the advancement in multilingual Large Language Models (mLLMs), emerging possibilities have been created to address the challenges of dialectal Automated Speech Recognition (ASR) and Machine Translation (MT). This study presents an end-to-end pipeline for converting dialectal Noakhali speech to standard Bangla speech. This investigation includes constructing a large-scale diverse dataset with dialectal speech signals that tailored the fine-tuning process in ASR and LLM for transcribing the dialect speech to dialect text and translating the dialect text to standard Bangla text. Our experiments demonstrated that fine-tuning the Whisper ASR model achieved a CER of 0.8% and WER of 1.5%, while the BanglaT5 model attained a BLEU score of 41.6% for dialect-to-standard text translation.
Authors: Quan Ze Chen, K. J. Kevin Feng, Chan Young Park, Amy X. Zhang
Abstract: Alignment of large language models (LLMs) to societal values should account for pluralistic values from diverse groups. One technique uses in-context learning for inference-time alignment, but only considers similarity when drawing few-shot examples, not accounting for cross-group differences in value prioritization. We propose SPICA, a framework for pluralistic alignment that accounts for group-level differences during in-context example retrieval. SPICA introduces three designs to facilitate pluralistic alignment: scenario banks, group-informed metrics, and in-context alignment prompts. From an evaluation of SPICA on an alignment task collecting inputs from four demographic groups ($n = 544$), our metrics retrieve in-context examples that more closely match observed preferences, with the best prompt configuration using multiple contrastive responses to demonstrate examples. In an end-to-end evaluation ($n = 80$), we observe that SPICA-aligned models are higher rated than a baseline similarity-only retrieval approach, with groups seeing up to a +0.16 point improvement on a 5 point scale. Additionally, gains from SPICA were more uniform, with all groups benefiting from alignment rather than only some. Finally, we find that while a group-agnostic approach can effectively align to aggregated values, it is not most suited for aligning to divergent groups.
Authors: Sizhe Wang, Yongqi Tong, Hengyuan Zhang, Dawei Li, Xin Zhang, Tianlong Chen
Abstract: Reinforcement Learning with Human Feedback (RLHF) is the key to the success of large language models (LLMs) in recent years. In this work, we first introduce the concepts of knowledge breadth and knowledge depth, which measure the comprehensiveness and depth of an LLM or knowledge source respectively. We reveal that the imbalance in the number of prompts and responses can lead to a potential disparity in breadth and depth learning within alignment tuning datasets by showing that even a simple uniform method for balancing the number of instructions and responses can lead to significant improvements. Building on this, we further propose Balanced Preference Optimization (BPO), designed to dynamically augment the knowledge depth of each sample. BPO is motivated by the observation that the usefulness of knowledge varies across samples, necessitating tailored learning of knowledge depth. To achieve this, we introduce gradient-based clustering, estimating the knowledge informativeness and usefulness of each augmented sample based on the model's optimization direction. Our experimental results across various benchmarks demonstrate that BPO outperforms other baseline methods in alignment tuning while maintaining training efficiency. Furthermore, we conduct a detailed analysis of each component of BPO, providing guidelines for future research in preference data optimization.
Authors: Yufei Guo, Muzhe Guo, Juntao Su, Zhou Yang, Mengqiu Zhu, Hongfei Li, Mengyang Qiu, Shuo Shuo Liu
Abstract: Large Language Models (LLMs) have revolutionized natural language processing, but their susceptibility to biases poses significant challenges. This comprehensive review examines the landscape of bias in LLMs, from its origins to current mitigation strategies. We categorize biases as intrinsic and extrinsic, analyzing their manifestations in various NLP tasks. The review critically assesses a range of bias evaluation methods, including data-level, model-level, and output-level approaches, providing researchers with a robust toolkit for bias detection. We further explore mitigation strategies, categorizing them into pre-model, intra-model, and post-model techniques, highlighting their effectiveness and limitations. Ethical and legal implications of biased LLMs are discussed, emphasizing potential harms in real-world applications such as healthcare and criminal justice. By synthesizing current knowledge on bias in LLMs, this review contributes to the ongoing effort to develop fair and responsible AI systems. Our work serves as a comprehensive resource for researchers and practitioners working towards understanding, evaluating, and mitigating bias in LLMs, fostering the development of more equitable AI technologies.
Authors: Jisang Park, Minu Kim, DaYoung Hong, Jongha Lee
Abstract: Learners of a second language (L2) often unconsciously substitute unfamiliar L2 phonemes with similar phonemes from their native language (L1), even though native speakers of the L2 perceive these sounds as distinct and non-interchangeable. This phonemic substitution leads to deviations from the standard phonological patterns of the L2, creating challenges for learners in acquiring accurate L2 pronunciation. To address this, we propose Inter-linguistic Phonetic Composition (IPC), a novel computational method designed to minimize incorrect phonological transfer by reconstructing L2 phonemes as composite sounds derived from multiple L1 phonemes. Tests with two automatic speech recognition models demonstrated that when L2 speakers produced IPC-generated composite sounds, the recognition rate of target L2 phonemes improved by 20% compared to when their pronunciation was influenced by original phonological transfer patterns. The improvement was observed within a relatively shorter time frame, demonstrating rapid acquisition of the composite sound.
Authors: Wenke Huang, Jian Liang, Zekun Shi, Didi Zhu, Guancheng Wan, He Li, Bo Du, Dacheng Tao, Mang Ye
Abstract: Multimodal Large Language Model (MLLM) have demonstrated strong generalization capabilities across diverse distributions and tasks, largely due to extensive pre-training datasets. Fine-tuning MLLM has become a common practice to improve performance on specific downstream tasks. However, during fine-tuning, MLLM often faces the risk of forgetting knowledge acquired during pre-training, which can result in a decline in generalization abilities. To balance the trade-off between generalization and specialization, we propose measuring the parameter importance for both pre-trained and fine-tuning distributions, based on frozen pre-trained weight magnitude and accumulated fine-tuning gradient values. We further apply an importance-aware weight allocation strategy, selectively updating relatively important parameters for downstream tasks. We conduct empirical evaluations on both image captioning and visual question-answering tasks using various MLLM architectures. The comprehensive experimental analysis demonstrates the effectiveness of the proposed solution, highlighting the efficiency of the crucial modules in enhancing downstream specialization performance while mitigating generalization degradation in MLLM Fine-Tuning.
Authors: Mika H\"am\"al\"ainen, Jack Rueter, Khalid Alnajjar
Abstract: We present a novel digital humanities method for representing our Twitch chatters as user embeddings created by a large language model (LLM). We cluster these embeddings automatically using affinity propagation and further narrow this clustering down through manual analysis. We analyze the chat of one stream by each Twitch streamer: SmallAnt, DougDoug and PointCrow. Our findings suggest that each streamer has their own type of chatters, however two categories emerge for all of the streamers: supportive viewers and emoji and reaction senders. Repetitive message spammers is a shared chatter category for two of the streamers.
Authors: Zeping Yu, Sophia Ananiadou
Abstract: Understanding the mechanisms behind Large Language Models (LLMs) is crucial for designing improved models and strategies. While recent studies have yielded valuable insights into the mechanisms of textual LLMs, the mechanisms of Multi-modal Large Language Models (MLLMs) remain underexplored. In this paper, we apply mechanistic interpretability methods to analyze the visual question answering (VQA) mechanisms in the first MLLM, Llava. We compare the mechanisms between VQA and textual QA (TQA) in color answering tasks and find that: a) VQA exhibits a mechanism similar to the in-context learning mechanism observed in TQA; b) the visual features exhibit significant interpretability when projecting the visual embeddings into the embedding space; and c) Llava enhances the existing capabilities of the corresponding textual LLM Vicuna during visual instruction tuning. Based on these findings, we develop an interpretability tool to help users and researchers identify important visual locations for final predictions, aiding in the understanding of visual hallucination. Our method demonstrates faster and more effective results compared to existing interpretability approaches. Code: \url{https://github.com/zepingyu0512/llava-mechanism}
Authors: Fahim Faisal, Md Mushfiqur Rahman, Antonios Anastasopoulos
Abstract: There has been little systematic study on how dialectal differences affect toxicity detection by modern LLMs. Furthermore, although using LLMs as evaluators ("LLM-as-a-judge") is a growing research area, their sensitivity to dialectal nuances is still underexplored and requires more focused attention. In this paper, we address these gaps through a comprehensive toxicity evaluation of LLMs across diverse dialects. We create a multi-dialect dataset through synthetic transformations and human-assisted translations, covering 10 language clusters and 60 varieties. We then evaluated three LLMs on their ability to assess toxicity across multilingual, dialectal, and LLM-human consistency. Our findings show that LLMs are sensitive in handling both multilingual and dialectal variations. However, if we have to rank the consistency, the weakest area is LLM-human agreement, followed by dialectal consistency. Code repository: \url{https://github.com/ffaisal93/dialect_toxicity_llm_judge}
URLs: https://github.com/ffaisal93/dialect_toxicity_llm_judge
Authors: Da-Chen Lian, Shu-Kai Hsieh
Abstract: This study aims to fill the gap by constructing a topic-aware comparable corpus of Mainland Chinese Mandarin and Taiwanese Mandarin from the social media in Mainland China and Taiwan, respectively. Using Dcard for Taiwanese Mandarin and Sina Weibo for Mainland Chinese, we create a comparable corpus that updates regularly and reflects modern language use on social media.
Authors: Sibo Wei, Xueping Peng, Yi-fei Wang, Jiasheng Si, Weiyu Zhang, Wenpeng Lu, Xiaoming Wu, Yinglong Wang
Abstract: The rise of large language models (LLMs) has driven significant progress in medical applications, including traditional Chinese medicine (TCM). However, current medical LLMs struggle with TCM diagnosis and syndrome differentiation due to substantial differences between TCM and modern medical theory, and the scarcity of specialized, high-quality corpora. This paper addresses these challenges by proposing BianCang, a TCM-specific LLM, using a two-stage training process that first injects domain-specific knowledge and then aligns it through targeted stimulation. To enhance diagnostic and differentiation capabilities, we constructed pre-training corpora, instruction-aligned datasets based on real hospital records, and the ChP-TCM dataset derived from the Pharmacopoeia of the People's Republic of China. We compiled extensive TCM and medical corpora for continuous pre-training and supervised fine-tuning, building a comprehensive dataset to refine the model's understanding of TCM. Evaluations across 11 test sets involving 29 models and 4 tasks demonstrate the effectiveness of BianCang, offering valuable insights for future research. Code, datasets, and models are available at https://github.com/QLU-NLP/BianCang.
Authors: Bin Xu, Yiguan Lin, Yinghao Li, YangGao
Abstract: Large language models demonstrate exceptional performance in simple code generation tasks but still face challenges in tackling complex problems. These challenges may stem from insufficient reasoning and problem decomposition capabilities. To address this issue, we propose a reasoning-augmented data generation process, SRA-MCTS, which guides the model to autonomously generate high-quality intermediate reasoning paths. This creates a positive feedback loop, enabling continuous improvement. Our method operates entirely through the model itself without requiring additional supervision. By synthesizing natural language reasoning paths and translating them into executable code, the approach ensures analytical accuracy and enhances the success rate in solving complex tasks. Experimental results show that, even without additional supervisory signals, our method achieves performance improvements across different model scales, demonstrating the significant potential of self-improvement in small models. Furthermore, the method remains robust when traditional Chain-of-Thought (CoT) approaches exhibit performance degradation, with notable improvements observed in diversity metrics such as pass@10. We encourage further exploration of reasoning processes within training data to enhance the ability of language models to address complex problems.
Authors: Ofir Zafrir, Igor Margulis, Dorin Shteyman, Guy Boudoukh
Abstract: Speculative Decoding has gained popularity as an effective technique for accelerating the auto-regressive inference process of Large Language Models (LLMs). However, Speculative Decoding entirely relies on the availability of efficient draft models, which are often lacking for many existing language models due to a stringent constraint of vocabulary incompatibility. In this work we introduce FastDraft, a novel and efficient approach for pre-training and aligning a draft model to any large language model by incorporating efficient pre-training, followed by fine-tuning over synthetic datasets generated by the target model. We demonstrate FastDraft by training two highly parameter efficient drafts for the popular Phi-3-mini and Llama-3.1-8B models. Using FastDraft, we were able to produce a draft with approximately 10 billion tokens on a single server with 8 Intel$^\circledR$ Gaudi$^\circledR$ 2 accelerators in under 24 hours. Our results show that the draft model achieves impressive results in key metrics of acceptance rate, block efficiency and up to 3x memory bound speed up when evaluated on code completion and up to 2x in summarization, text completion and instruction tasks. We validate our theoretical findings through benchmarking on the latest Intel$^\circledR$ Core$^{\tiny \text{TM}}$ Ultra, achieving a wall-clock time speedup of up to 2x, indicating a significant reduction in runtime. Due to its high quality, FastDraft unlocks large language models inference on AI-PC and other edge-devices.
Authors: Xiaoliang Luo, Michael Ramscar, Bradley C. Love
Abstract: The impressive performance of large language models (LLMs) has led to their consideration as models of human language processing. Instead, we suggest that the success of LLMs arises from the flexibility of the transformer learning architecture. To evaluate this conjecture, we trained LLMs on scientific texts that were either in a forward or backward format. Despite backward text being inconsistent with the structure of human languages, we found that LLMs performed equally well in either format on a neuroscience benchmark, eclipsing human expert performance for both forward and backward orders. Our results are consistent with the success of transformers across diverse domains, such as weather prediction and protein design. This widespread success is attributable to LLM's ability to extract predictive patterns from any sufficiently structured input. Given their generality, we suggest caution in interpreting LLM's success in linguistic tasks as evidence for human-like mechanisms.
Authors: Shaolin Zhu, Supryadi, Shaoyang Xu, Haoran Sun, Leiyu Pan, Menglong Cui, Jiangcun Du, Renren Jin, Ant\'onio Branco, Deyi Xiong
Abstract: This paper provides a comprehensive survey of the latest research on multilingual large language models (MLLMs). MLLMs not only are able to understand and generate language across linguistic boundaries, but also represent an important advancement in artificial intelligence. We first discuss the architecture and pre-training objectives of MLLMs, highlighting the key components and methodologies that contribute to their multilingual capabilities. We then discuss the construction of multilingual pre-training and alignment datasets, underscoring the importance of data quality and diversity in enhancing MLLM performance. An important focus of this survey is on the evaluation of MLLMs. We present a detailed taxonomy and roadmap covering the assessment of MLLMs' cross-lingual knowledge, reasoning, alignment with human values, safety, interpretability and specialized applications. Specifically, we extensively discuss multilingual evaluation benchmarks and datasets, and explore the use of LLMs themselves as multilingual evaluators. To enhance MLLMs from black to white boxes, we also address the interpretability of multilingual capabilities, cross-lingual transfer and language bias within these models. Finally, we provide a comprehensive review of real-world applications of MLLMs across diverse domains, including biology, medicine, computer science, mathematics and law. We showcase how these models have driven innovation and improvements in these specialized fields while also highlighting the challenges and opportunities in deploying MLLMs within diverse language communities and application scenarios.We listed the paper related in this survey and publicly available at https://github.com/tjunlp-lab/Awesome-Multilingual-LLMs-Papers .
URLs: https://github.com/tjunlp-lab/Awesome-Multilingual-LLMs-Papers
Authors: Tomas Horych, Christoph Mandl, Terry Ruas, Andre Greiner-Petter, Bela Gipp, Akiko Aizawa, Timo Spinde
Abstract: High annotation costs from hiring or crowdsourcing complicate the creation of large, high-quality datasets needed for training reliable text classifiers. Recent research suggests using Large Language Models (LLMs) to automate the annotation process, reducing these costs while maintaining data quality. LLMs have shown promising results in annotating downstream tasks like hate speech detection and political framing. Building on the success in these areas, this study investigates whether LLMs are viable for annotating the complex task of media bias detection and whether a downstream media bias classifier can be trained on such data. We create annolexical, the first large-scale dataset for media bias classification with over 48000 synthetically annotated examples. Our classifier, fine-tuned on this dataset, surpasses all of the annotator LLMs by 5-9 percent in Matthews Correlation Coefficient (MCC) and performs close to or outperforms the model trained on human-labeled data when evaluated on two media bias benchmark datasets (BABE and BASIL). This study demonstrates how our approach significantly reduces the cost of dataset creation in the media bias domain and, by extension, the development of classifiers, while our subsequent behavioral stress-testing reveals some of its current limitations and trade-offs.
Authors: Jan Pfister, Julia Wunderle, Andreas Hotho
Abstract: We create two German-only decoder models, LL\"aMmlein 120M and 1B, transparently from scratch and publish them, along with the training data, for the German NLP research community to use. The model training involved several key steps, including extensive data preprocessing, the creation of a custom German tokenizer, the training itself, as well as the evaluation of the final models on various benchmarks. Throughout the training process, multiple checkpoints were saved and analyzed using the SuperGLEBer benchmark to monitor the models' learning dynamics. Compared to state-of-the-art models on the SuperGLEBer benchmark, both LL\"aMmlein models performed competitively, consistently matching or surpassing models with similar parameter sizes. The results show that the models' quality scales with size as expected, but performance improvements on some tasks plateaued early, offering valuable insights into resource allocation for future model development.
Authors: Martin Andrews
Abstract: Excellent progress has been made recently in solving ARC Challenge problems. However, it seems that new techniques may be required to push beyond 60% accuracy. Even commercial Large Language Models (LLMs) struggle to 'understand' many of the problems (when given the input and output grids), which makes discovering solutions by LLM-lead program search somewhat futile. In this work, LLM 'understanding' is attempted from a stronger starting position : An LLM is given complete solutions to tasks in code, and then asked to explain how the task is being solved at various levels of abstraction. Specifically, the LLM was given code solutions implemented in arc-dsl-llm (an LLM-legible version of Hodel's arc-dsl to obtain: (a) commented code; (b) code refactored into reusable functional chunks; (c) problem solution steps; and (d) high-level problem-solving tactics. We demonstrate that 'Sparks of Abstraction' can be extracted from the LLM output - in a form that could be used in downstream tasks with Local LLMs eligible to enter the ARC Prize. Both the arc-dsl-llm DSL framework (with the re-engineered solutions) and the Gemini LLM-generated data (along with the generation code) are made Open Source.
Authors: Yingjie Zhou, Zicheng Zhang, Jiezhang Cao, Jun Jia, Yanwei Jiang, Farong Wen, Xiaohong Liu, Xiongkuo Min, Guangtao Zhai
Abstract: Artificial Intelligence (AI) has demonstrated significant capabilities in various fields, and in areas such as human-computer interaction (HCI), embodied intelligence, and the design and animation of virtual digital humans, both practitioners and users are increasingly concerned with AI's ability to understand and express emotion. Consequently, the question of whether AI can accurately interpret human emotions remains a critical challenge. To date, two primary classes of AI models have been involved in human emotion analysis: generative models and Multimodal Large Language Models (MLLMs). To assess the emotional capabilities of these two classes of models, this study introduces MEMO-Bench, a comprehensive benchmark consisting of 7,145 portraits, each depicting one of six different emotions, generated by 12 Text-to-Image (T2I) models. Unlike previous works, MEMO-Bench provides a framework for evaluating both T2I models and MLLMs in the context of sentiment analysis. Additionally, a progressive evaluation approach is employed, moving from coarse-grained to fine-grained metrics, to offer a more detailed and comprehensive assessment of the sentiment analysis capabilities of MLLMs. The experimental results demonstrate that existing T2I models are more effective at generating positive emotions than negative ones. Meanwhile, although MLLMs show a certain degree of effectiveness in distinguishing and recognizing human emotions, they fall short of human-level accuracy, particularly in fine-grained emotion analysis. The MEMO-Bench will be made publicly available to support further research in this area.
Authors: Son T. Luu, Hiep Nguyen, Trung Vo, Le-Minh Nguyen
Abstract: In this paper, we propose ZeFaV - a zero-shot based fact-checking verification framework to enhance the performance on fact verification task of large language models by leveraging the in-context learning ability of large language models to extract the relations among the entities within a claim, re-organized the information from the evidence in a relationally logical form, and combine the above information with the original evidence to generate the context from which our fact-checking model provide verdicts for the input claims. We conducted empirical experiments to evaluate our approach on two multi-hop fact-checking datasets including HoVer and FEVEROUS, and achieved potential results results comparable to other state-of-the-art fact verification task methods.
Authors: Cameron Morin, Matti Marttinen Larsson
Abstract: Much linguistic research relies on annotated datasets of features extracted from text corpora, but the rapid quantitative growth of these corpora has created practical difficulties for linguists to manually annotate large data samples. In this paper, we present a replicable, supervised method that leverages large language models for assisting the linguist in grammatical annotation through prompt engineering, training, and evaluation. We introduce a methodological pipeline applied to the case study of formal variation in the English evaluative verb construction 'consider X (as) (to be) Y', based on the large language model Claude 3.5 Sonnet and corpus data from Davies' NOW and EnTenTen21 (SketchEngine). Overall, we reach a model accuracy of over 90% on our held-out test samples with only a small amount of training data, validating the method for the annotation of very large quantities of tokens of the construction in the future. We discuss the generalisability of our results for a wider range of case studies of grammatical constructions and grammatical variation and change, underlining the value of AI copilots as tools for future linguistic research.
Authors: Keer Lu, Keshi Zhao, Zheng Liang, Da Pan, Shusen Zhang, Xin Wu, Weipeng Chen, Zenan Zhou, Guosheng Dong, Bin Cui, Wentao Zhang
Abstract: Large Language Models (LLMs) exhibit remarkable capabilities in handling multiple tasks across domains due to their emergent properties. These capabilities are further augmented during the Supervised Fine-Tuning (SFT) phase. Despite their potential, existing work mainly focuses on domain-specific enhancements during fine-tuning, the challenge of which lies in catastrophic forgetting of knowledge across other domains. In this study, we introduce VersaTune, a novel data composition framework designed for enhancing LLMs' overall multi-ability performances during fine-tuning. We categorize knowledge into distinct domains including law, medicine, finance, science, code. We begin with detecting the distribution of domain-specific knowledge within the base model, followed by the composition of training data that aligns with the model's existing knowledge distribution. During the fine-tuning process, weights of different domains are dynamically adjusted based on their learnable potential and forgetting degree. Experimental results demonstrate that VersaTune achieves significant improvements in multi-domain performance, with a 35.21% enhancement in comprehensive multi-domain tasks. Additionally, in scenarios where specific domain optimization is required, VersaTune reduces the degradation of performance in other domains by 38.77%, without compromising the target domain's training efficacy.
Authors: Yungi Kim, Hyunsoo Ha, Seonghoon Yang, Sukyung Lee, Jihoo Kim, Chanjun Park
Abstract: Creating high-quality, large-scale datasets for large language models (LLMs) often relies on resource-intensive, GPU-accelerated models for quality filtering, making the process time-consuming and costly. This dependence on GPUs limits accessibility for organizations lacking significant computational infrastructure. To address this issue, we introduce the Lightweight, Purpose-driven (LP) Data Pipeline, a framework that operates entirely on CPUs to streamline the processes of dataset extraction, filtering, and curation. Based on our four core principles, the LP Data Pipeline significantly reduces preparation time and cost while maintaining high data quality. Importantly, our pipeline enables the creation of purpose-driven datasets tailored to specific domains and languages, enhancing the applicability of LLMs in specialized contexts. We anticipate that our pipeline will lower the barriers to LLM development, enabling a wide range of organizations to access LLMs more easily.
Authors: Peng Shu, Junhao Chen, Zhengliang Liu, Hui Wang, Zihao Wu, Tianyang Zhong, Yiwei Li, Huaqin Zhao, Hanqi Jiang, Yi Pan, Yifan Zhou, Constance Owl, Xiaoming Zhai, Ninghao Liu, Claudio Saunt, Tianming Liu
Abstract: Large Language Models (LLMs) have demonstrated remarkable success across a wide range of tasks and domains. However, their performance in low-resource language translation, particularly when translating into these languages, remains underexplored. This gap poses significant challenges, as linguistic barriers hinder the cultural preservation and development of minority communities. To address this issue, this paper introduces a novel retrieval-based method that enhances translation quality for low-resource languages by focusing on key terms, which involves translating keywords and retrieving corresponding examples from existing data. To evaluate the effectiveness of this method, we conducted experiments translating from English into three low-resource languages: Cherokee, a critically endangered indigenous language of North America; Tibetan, a historically and culturally significant language in Asia; and Manchu, a language with few remaining speakers. Our comparison with the zero-shot performance of GPT-4o and LLaMA 3.1 405B, highlights the significant challenges these models face when translating into low-resource languages. In contrast, our retrieval-based method shows promise in improving both word-level accuracy and overall semantic understanding by leveraging existing resources more effectively.
Authors: Han Cao, Zhaoyang Zhang, Xiangtian Li, Chufan Wu, Hansong Zhang, Wenqing Zhang
Abstract: Knowledge-aware sequence to sequence generation tasks such as document question answering and abstract summarization typically requires two types of knowledge: encoded parametric knowledge and retrieved contextual information. Previous work show improper correlation between parametric knowledge and answers in the training set could cause the model ignore input information at test time, resulting in un-desirable model behaviour such as over-stability and hallucination. In this work, we argue that hallucination could be mitigated via explicit correlation between input source and generated content. We focus on a typical example of hallucination, entity-based knowledge conflicts in question answering, where correlation of entities and their description at training time hinders model behaviour during inference.
Authors: Sreeram Vennam, David Valente, David Herel, Ponnurangam Kumaraguru
Abstract: Thinking Tokens (TT) have been proposed as an unsupervised method to facilitate reasoning in language models. However, despite their conceptual appeal, our findings show that TTs marginally improves performance and consistently underperforms compared to Chain-of-Thought (CoT) reasoning across multiple benchmarks. We hypothesize that this underperformance stems from the reliance on a single embedding for TTs, which results in inconsistent learning signals and introduces noisy gradients. This paper provides a comprehensive empirical analysis to validate this hypothesis and discusses the implications for future research on unsupervised reasoning in LLMs.
Authors: Zixiong Wang, Gaoyang Liu, Yang Yang, Chen Wang
Abstract: Recent advances in Large Language Models (LLMs) have enabled them to overcome their context window limitations, and demonstrate exceptional retrieval and reasoning capacities on longer context. Quesion-answering systems augmented with Long-Context Language Models (LCLMs) can automatically search massive external data and incorporate it into their contexts, enabling faithful predictions and reducing issues such as hallucinations and knowledge staleness. Existing studies targeting LCLMs mainly concentrate on addressing the so-called lost-in-the-middle problem or improving the inference effiencicy, leaving their privacy risks largely unexplored. In this paper, we aim to bridge this gap and argue that integrating all information into the long context makes it a repository of sensitive information, which often contains private data such as medical records or personal identities. We further investigate the membership privacy within LCLMs external context, with the aim of determining whether a given document or sequence is included in the LCLMs context. Our basic idea is that if a document lies in the context, it will exhibit a low generation loss or a high degree of semantic similarity to the contents generated by LCLMs. We for the first time propose six membership inference attack (MIA) strategies tailored for LCLMs and conduct extensive experiments on various popular models. Empirical results demonstrate that our attacks can accurately infer membership status in most cases, e.g., 90.66% attack F1-score on Multi-document QA datasets with LongChat-7b-v1.5-32k, highlighting significant risks of membership leakage within LCLMs input contexts. Furthermore, we examine the underlying reasons why LCLMs are susceptible to revealing such membership information.
Authors: Jingxuan Li, Yuning Yang, Shengqi Yang, Yizhou Zhao, Ying Nian Wu
Abstract: The rapid advancement of Vision-Language Models (VLMs) has expanded multimodal applications, yet evaluations often focus on basic tasks like object recognition, overlooking abstract aspects such as personalities and values. To address this gap, we introduce Value-Spectrum, a visual question-answering benchmark aimed at assessing VLMs based on Schwartz's value dimensions, which capture core values guiding people's beliefs and actions across cultures. We constructed a vectorized database of over 50,000 short videos sourced from TikTok, YouTube Shorts, and Instagram Reels, covering multiple months and a wide array of topics such as family, health, hobbies, society, and technology. We also developed a VLM agent pipeline to automate video browsing and analysis. Benchmarking representative VLMs on Value-Spectrum reveals significant differences in their responses to value-oriented content, with most models exhibiting a preference for hedonistic topics. Beyond identifying natural preferences, we explored the ability of VLM agents to adopt specific personas when explicitly prompted, revealing insights into the models' adaptability in role-playing scenarios. These findings highlight the potential of Value-Spectrum as a comprehensive evaluation set for tracking VLM advancements in value-based tasks and for developing more sophisticated role-playing AI agents.
Authors: Chenhang Cui, Gelei Deng, An Zhang, Jingnan Zheng, Yicong Li, Lianli Gao, Tianwei Zhang, Tat-Seng Chua
Abstract: Recent advances in Large Vision-Language Models (LVLMs) have showcased strong reasoning abilities across multiple modalities, achieving significant breakthroughs in various real-world applications. Despite this great success, the safety guardrail of LVLMs may not cover the unforeseen domains introduced by the visual modality. Existing studies primarily focus on eliciting LVLMs to generate harmful responses via carefully crafted image-based jailbreaks designed to bypass alignment defenses. In this study, we reveal that a safe image can be exploited to achieve the same jailbreak consequence when combined with additional safe images and prompts. This stems from two fundamental properties of LVLMs: universal reasoning capabilities and safety snowball effect. Building on these insights, we propose Safety Snowball Agent (SSA), a novel agent-based framework leveraging agents' autonomous and tool-using abilities to jailbreak LVLMs. SSA operates through two principal stages: (1) initial response generation, where tools generate or retrieve jailbreak images based on potential harmful intents, and (2) harmful snowballing, where refined subsequent prompts induce progressively harmful outputs. Our experiments demonstrate that \ours can use nearly any image to induce LVLMs to produce unsafe content, achieving high success jailbreaking rates against the latest LVLMs. Unlike prior works that exploit alignment flaws, \ours leverages the inherent properties of LVLMs, presenting a profound challenge for enforcing safety in generative multimodal systems. Our code is avaliable at \url{https://github.com/gzcch/Safety_Snowball_Agent}.
Authors: Viktoriia Chekalina, Anton Razzigaev, Elizaveta Goncharova, Andrey Kuznetsov
Abstract: In this paper we present an approach to reduce hallucinations in Large Language Models (LLMs) by incorporating Knowledge Graphs (KGs) as an additional modality. Our method involves transforming input text into a set of KG embeddings and using an adapter to integrate these embeddings into the language model space, without relying on external retrieval processes. To facilitate this, we created WikiEntities, a dataset containing over 3 million Wikipedia texts annotated with entities from Wikidata and their corresponding embeddings from PyTorch-BigGraph. This dataset serves as a valuable resource for training Entity Linking models and adapting the described method to various LLMs using specialized adapters. Our method does not require fine-tuning of the language models themselves; instead, we only train the adapter. This ensures that the model's performance on other tasks is not affected. We trained an adapter for the Mistral 7B, LLaMA 2-7B (chat), and LLaMA 3-8B (instruct) models using this dataset and demonstrated that our approach improves performance on the HaluEval, True-False benchmarks and FEVER dataset. The results indicate that incorporating KGs as a new modality can effectively reduce hallucinations and improve the factual accuracy of language models, all without the need for external retrieval.
Authors: Ziyi Yang, Zaibin Zhang, Zirui Zheng, Yuxian Jiang, Ziyue Gan, Zhiyu Wang, Zijian Ling, Jinsong Chen, Martz Ma, Bowen Dong, Prateek Gupta, Shuyue Hu, Zhenfei Yin, Guohao Li, Xu Jia, Lijun Wang, Bernard Ghanem, Huchuan Lu, Wanli Ouyang, Yu Qiao, Philip Torr, Jing Shao
Abstract: There has been a growing interest in enhancing rule-based agent-based models (ABMs) for social media platforms (\emph{i.e.}, X, Reddit) with more realistic large language model (LLM) agents, thereby allowing for a more nuanced study of complex systems. As a result, several LLM-based ABMs have been proposed in the past year. While they hold promise, each simulator is specifically designed to study a particular scenario, making it time-consuming and resource-intensive to explore other phenomena using the same ABM. Additionally, these models simulate only a limited number of agents, whereas real-world social media platforms involve millions of users. To this end, we propose OASIS, a generalizable and scalable social media simulator. OASIS is designed based on real-world social media platforms, incorporating dynamically updated environments (\emph{i.e.}, dynamic social networks and post information), diverse action spaces (\emph{i.e.}, following, commenting), and recommendation systems (\emph{i.e.}, interest-based and hot-score-based). Additionally, OASIS supports large-scale user simulations, capable of modeling up to one million users. With these features, OASIS can be easily extended to different social media platforms to study large-scale group phenomena and behaviors. We replicate various social phenomena, including information spreading, group polarization, and herd effects across X and Reddit platforms. Moreover, we provide observations of social phenomena at different agent group scales. We observe that the larger agent group scale leads to more enhanced group dynamics and more diverse and helpful agents' opinions. These findings demonstrate OASIS's potential as a powerful tool for studying complex systems in digital environments.
Authors: Duzhen Zhang, Yahan Yu, Chenxing Li, Jiahua Dong, Dong Yu
Abstract: Federated Named Entity Recognition (FNER) boosts model training within each local client by aggregating the model updates of decentralized local clients, without sharing their private data. However, existing FNER methods assume fixed entity types and local clients in advance, leading to their ineffectiveness in practical applications. In a more realistic scenario, local clients receive new entity types continuously, while new local clients collecting novel data may irregularly join the global FNER training. This challenging setup, referred to here as Federated Incremental NER, renders the global model suffering from heterogeneous forgetting of old entity types from both intra-client and inter-client perspectives. To overcome these challenges, we propose a Local-Global Forgetting Defense (LGFD) model. Specifically, to address intra-client forgetting, we develop a structural knowledge distillation loss to retain the latent space's feature structure and a pseudo-label-guided inter-type contrastive loss to enhance discriminative capability over different entity types, effectively preserving previously learned knowledge within local clients. To tackle inter-client forgetting, we propose a task switching monitor that can automatically identify new entity types under privacy protection and store the latest old global model for knowledge distillation and pseudo-labeling. Experiments demonstrate significant improvement of our LGFD model over comparison methods.
Authors: Leon Kopitar, Primoz Kocbek, Lucija Gosak, Gregor Stiglic
Abstract: This review examines the development of abstractive NLP-based text summarization approaches and compares them to existing techniques for extractive summarization. A brief history of text summarization from the 1950s to the introduction of pre-trained language models such as Bidirectional Encoder Representations from Transformer (BERT) and Generative Pre-training Transformers (GPT) are presented. In total, 60 studies were identified in PubMed and Web of Science, of which 29 were excluded and 24 were read and evaluated for eligibility, resulting in the use of seven studies for further analysis. This chapter also includes a section with examples including an example of a comparison between GPT-3 and state-of-the-art GPT-4 solutions in scientific text summarisation. Natural language processing has not yet reached its full potential in the generation of brief textual summaries. As there are acknowledged concerns that must be addressed, we can expect gradual introduction of such models in practise.
Authors: Jinhao Jiang, Zhipeng Chen, Yingqian Min, Jie Chen, Xiaoxue Cheng, Jiapeng Wang, Yiru Tang, Haoxiang Sun, Jia Deng, Wayne Xin Zhao, Zheng Liu, Dong Yan, Jian Xie, Zhongyuan Wang, Ji-Rong Wen
Abstract: Recently, test-time scaling has garnered significant attention from the research community, largely due to the substantial advancements of the o1 model released by OpenAI. By allocating more computational resources during the inference phase, large language models~(LLMs) can extensively explore the solution space by generating more thought tokens or diverse solutions, thereby producing more accurate responses. However, developing an o1-like reasoning approach is challenging, and researchers have been making various attempts to advance this open area of research. In this paper, we present a preliminary exploration into enhancing the reasoning abilities of LLMs through reward-guided tree search algorithms. This framework is implemented by integrating the policy model, reward model, and search algorithm. It is primarily constructed around a tree search algorithm, where the policy model navigates a dynamically expanding tree guided by a specially trained reward model. We thoroughly explore various design considerations necessary for implementing this framework and provide a detailed report of the technical aspects. To assess the effectiveness of our approach, we focus on mathematical reasoning tasks and conduct extensive evaluations on four challenging datasets, significantly enhancing the reasoning abilities of LLMs.
Authors: Tao Fan, Yan Kang, Guoqiang Ma, Lixin Fan, Kai Chen, Qiang Yang
Abstract: By adapting Large Language Models (LLMs) to domain-specific tasks or enriching them with domain-specific knowledge, we can fully harness the capabilities of LLMs. Nonetheless, a gap persists in achieving simultaneous mutual enhancement between the server's LLM and the downstream clients' Small Language Models (SLMs). To address this, we propose FedCoLLM, a novel and parameter-efficient federated framework designed for co-tuning LLMs and SLMs. This approach is aimed at adaptively transferring server-side LLMs knowledge to clients' SLMs while simultaneously enriching the LLMs with domain insights from the clients. To accomplish this, FedCoLLM utilizes lightweight adapters in conjunction with SLMs, facilitating knowledge exchange between server and clients in a manner that respects data privacy while also minimizing computational and communication overhead. Our evaluation of FedCoLLM, utilizing various public LLMs and SLMs across a range of NLP text generation tasks, reveals that the performance of clients' SLMs experiences notable improvements with the assistance of the LLMs. Simultaneously, the LLMs enhanced via FedCoLLM achieves comparable performance to that obtained through direct fine-tuning on clients' data.
Authors: Allison Huang, Yulu Niki Pi, Carlos Mougan
Abstract: We explore how large language models (LLMs) can be influenced by prompting them to alter their initial decisions and align them with established ethical frameworks. Our study is based on two experiments designed to assess the susceptibility of LLMs to moral persuasion. In the first experiment, we examine the susceptibility to moral ambiguity by evaluating a Base Agent LLM on morally ambiguous scenarios and observing how a Persuader Agent attempts to modify the Base Agent's initial decisions. The second experiment evaluates the susceptibility of LLMs to align with predefined ethical frameworks by prompting them to adopt specific value alignments rooted in established philosophical theories. The results demonstrate that LLMs can indeed be persuaded in morally charged scenarios, with the success of persuasion depending on factors such as the model used, the complexity of the scenario, and the conversation length. Notably, LLMs of distinct sizes but from the same company produced markedly different outcomes, highlighting the variability in their susceptibility to ethical persuasion.
Authors: German Gritsai, Anastasia Voznyuk, Ildar Khabutdinov, Andrey Grabovoy
Abstract: The paper describes a system designed by Advacheck team to recognise machine-generated and human-written texts in the monolingual subtask of GenAI Detection Task 1 competition. Our developed system is a multi-task architecture with shared Transformer Encoder between several classification heads. One head is responsible for binary classification between human-written and machine-generated texts, while the other heads are auxiliary multiclass classifiers for texts of different domains from particular datasets. As multiclass heads were trained to distinguish the domains presented in the data, they provide a better understanding of the samples. This approach led us to achieve the first place in the official ranking with 83.07% macro F1-score on the test set and bypass the baseline by 10%. We further study obtained system through ablation, error and representation analyses, finding that multi-task learning outperforms single-task mode and simultaneous tasks form a cluster structure in embeddings space.
Authors: Zishuo Feng, Feng Cao
Abstract: The task of converting Hanyu Pinyin abbreviations to Chinese characters represents a significant branch within the domain of Chinese Spelling Correction (CSC). This task is typically one of text-length alignment, however, due to the limited informational content in pinyin abbreviations, achieving accurate conversion is challenging. In this paper, we propose CNMBert which stands for zh-CN Pinyin Multi-mask Bert Model as a solution to this issue. CNMBert surpasses few-shot GPT models, achieving a 59.63% MRR on a 10,424-sample Hanyu Pinyin abbreviation test dataset.
Authors: Shengkun Tang, Liqun Ma, Haonan Li, Mingjie Sun, Zhiqiang Shen
Abstract: The typical selective state-space model (SSM) of Mamba addresses several limitations of Transformers, such as quadratic computational complexity with sequence length and significant inference-time memory requirements due to the key-value cache. However, the growing size of Mamba models continues to pose training and deployment challenges and raises environmental concerns due to considerable energy consumption. In this work, we introduce Bi-Mamba, a scalable and powerful 1-bit Mamba architecture designed for more efficient large language models with multiple sizes across 780M, 1.3B, and 2.7B. Bi-Mamba models are trained from scratch on data volume as regular LLM pertaining using an autoregressive distillation loss. Extensive experimental results on language modeling demonstrate that Bi-Mamba achieves performance comparable to its full-precision counterparts (e.g., FP16 or BF16) and much better accuracy than post-training-binarization (PTB) Mamba baselines, while significantly reducing memory footprint and energy consumption compared to the original Mamba model. Our study pioneers a new linear computational complexity LLM framework under low-bit representation and facilitates the future design of specialized hardware tailored for efficient 1-bit Mamba-based LLMs.
Authors: Edoardo Sebastiano De Duro, Enrique Taietta, Riccardo Improta, Massimo Stella
Abstract: Machine psychology aims to reconstruct the mindset of Large Language Models (LLMs), i.e. how these artificial intelligences perceive and associate ideas. This work introduces PhDGPT, a prompting framework and synthetic dataset that encapsulates the machine psychology of PhD researchers and professors as perceived by OpenAI's GPT-3.5. The dataset consists of 756,000 datapoints, counting 300 iterations repeated across 15 academic events, 2 biological genders, 2 career levels and 42 unique item responses of the Depression, Anxiety, and Stress Scale (DASS-42). PhDGPT integrates these psychometric scores with their explanations in plain language. This synergy of scores and texts offers a dual, comprehensive perspective on the emotional well-being of simulated academics, e.g. male/female PhD students or professors. By combining network psychometrics and psycholinguistic dimensions, this study identifies several similarities and distinctions between human and LLM data. The psychometric networks of simulated male professors do not differ between physical and emotional anxiety subscales, unlike humans. Other LLMs' personification can reconstruct human DASS factors with a purity up to 80%. Furthemore, LLM-generated personifications across different scenarios are found to elicit explanations lower in concreteness and imageability in items coding for anxiety, in agreement with past studies about human psychology. Our findings indicate an advanced yet incomplete ability for LLMs to reproduce the complexity of human psychometric data, unveiling convenient advantages and limitations in using LLMs to replace human participants. PhDGPT also intriguingly capture the ability for LLMs to adapt and change language patterns according to prompted mental distress contextual features, opening new quantitative opportunities for assessing the machine psychology of these artificial intelligences.
Authors: Rongxin Ouyang, Kokil Jaidka, Subhayan Mukerjee, Guangyu Cui
Abstract: The prevalence of multi-modal content on social media complicates automated moderation strategies. This calls for an enhancement in multi-modal classification and a deeper understanding of understated meanings in images and memes. Although previous efforts have aimed at improving model performance through fine-tuning, few have explored an end-to-end optimization pipeline that accounts for modalities, prompting, labeling, and fine-tuning. In this study, we propose an end-to-end conceptual framework for model optimization in complex tasks. Experiments support the efficacy of this traditional yet novel framework, achieving the highest accuracy and AUROC. Ablation experiments demonstrate that isolated optimizations are not ineffective on their own.
Authors: G. Thomas Hudson, Dean Slack, Thomas Winterbottom, Jamie Sterling, Chenghao Xiao, Junjie Shentu, Noura Al Moubayed
Abstract: Multimodal learning, which involves integrating information from various modalities such as text, images, audio, and video, is pivotal for numerous complex tasks like visual question answering, cross-modal retrieval, and caption generation. Traditional approaches rely on modality-specific encoders and late fusion techniques, which can hinder scalability and flexibility when adapting to new tasks or modalities. To address these limitations, we introduce a novel framework that extends the concept of task reformulation beyond natural language processing (NLP) to multimodal learning. We propose to reformulate diverse multimodal tasks into a unified next-frame prediction problem, allowing a single model to handle different modalities without modality-specific components. This method treats all inputs and outputs as sequential frames in a video, enabling seamless integration of modalities and effective knowledge transfer across tasks. Our approach is evaluated on a range of tasks, including text-to-text, image-to-text, video-to-video, video-to-text, and audio-to-text, demonstrating the model's ability to generalize across modalities with minimal adaptation. We show that task reformulation can significantly simplify multimodal model design across various tasks, laying the groundwork for more generalized multimodal foundation models.
Authors: Priyansh Bhatnagar, Linfeng Wen, Mingu Kang
Abstract: Extensive efforts have been made to boost the performance in the domain of language models by introducing various attention-based transformers. However, the inclusion of linear layers with large dimensions contributes to significant computational and memory overheads. The escalating computational demands of these models necessitate the development of various compression techniques to ensure their deployment on devices, particularly in resource-constrained environments. In this paper, we propose a novel compression methodology that dynamically determines the rank of each layer using a soft thresholding mechanism, which clips the singular values with a small magnitude in a differentiable form. This approach automates the decision-making process to identify the optimal degree of compression for each layer. We have successfully applied the proposed technique to attention-based architectures, including BERT for discriminative tasks and GPT2 and TinyLlama for generative tasks. Additionally, we have validated our method on Mamba, a recently proposed state-space model. Our experiments demonstrate that the proposed technique achieves a speed-up of 1.33X to 1.72X in the encoder/ decoder with a 50% reduction in total parameters.
Authors: Amrit Khera, Rajat Ghosh, Debojyoti Dutta
Abstract: LLM alignment ensures that large language models behave safely and effectively by aligning their outputs with human values, goals, and intentions. Aligning LLMs employ huge amounts of data, computation, and time. Moreover, curating data with human feedback is expensive and takes time. Recent research depicts the benefit of data engineering in the fine-tuning and pre-training paradigms to bring down such costs. However, alignment differs from the afore-mentioned paradigms and it is unclear if data efficient alignment is feasible. In this work, we first aim to understand how the performance of LLM alignment scales with data. We find out that LLM alignment performance follows an exponential plateau pattern which tapers off post a rapid initial increase. Based on this, we identify data subsampling as a viable method to reduce resources required for alignment. Further, we propose an information theory-based methodology for efficient alignment by identifying a small high quality subset thereby reducing the computation and time required by alignment. We evaluate the proposed methodology over multiple datasets and compare the results. We find that the model aligned using our proposed methodology outperforms other sampling methods and performs comparable to the model aligned with the full dataset while using less than 10% data, leading to greater than 90% savings in costs, resources, and faster LLM alignment.
Authors: Moshe Kimhi, Idan Kashani, Avi Mendelson, Chaim Baskin
Abstract: The widely used ReLU is favored for its hardware efficiency, {as the implementation at inference is a one bit sign case,} yet suffers from issues such as the ``dying ReLU'' problem, where during training, neurons fail to activate and constantly remain at zero, as highlighted by Lu et al. Traditional approaches to mitigate this issue often introduce more complex and less hardware-friendly activation functions. In this work, we propose a Hysteresis Rectified Linear Unit (HeLU), an efficient activation function designed to address the ``dying ReLU'' problem with minimal complexity. Unlike traditional activation functions with fixed thresholds for training and inference, HeLU employs a variable threshold that refines the backpropagation. This refined mechanism allows simpler activation functions to achieve competitive performance comparable to their more complex counterparts without introducing unnecessary complexity or requiring inductive biases. Empirical evaluations demonstrate that HeLU enhances model generalization across diverse datasets, offering a promising solution for efficient and effective inference suitable for a wide range of neural network architectures.
Authors: Yunsheng Ma, Burhaneddin Yaman, Xin Ye, Feng Tao, Abhirup Mallik, Ziran Wang, Liu Ren
Abstract: Bird's eye view (BEV)-based 3D perception plays a crucial role in autonomous driving applications. The rise of large language models has spurred interest in BEV-based captioning to understand object behavior in the surrounding environment. However, existing approaches treat perception and captioning as separate tasks, focusing on the performance of only one of the tasks and overlooking the potential benefits of multimodal alignment. To bridge this gap between modalities, we introduce MTA, a novel multimodal task alignment framework that boosts both BEV perception and captioning. MTA consists of two key components: (1) BEV-Language Alignment (BLA), a contextual learning mechanism that aligns the BEV scene representations with ground-truth language representations, and (2) Detection-Captioning Alignment (DCA), a cross-modal prompting mechanism that aligns detection and captioning outputs. MTA integrates into state-of-the-art baselines during training, adding no extra computational complexity at runtime. Extensive experiments on the nuScenes and TOD3Cap datasets show that MTA significantly outperforms state-of-the-art baselines, achieving a 4.9% improvement in perception and a 9.2% improvement in captioning. These results underscore the effectiveness of unified alignment in reconciling BEV-based perception and captioning.
Authors: Xudong Lu, Yinghao Chen, Cheng Chen, Hui Tan, Boheng Chen, Yina Xie, Rui Hu, Guanxin Tan, Renshou Wu, Yan Hu, Yi Zeng, Lei Wu, Liuyang Bian, Zhaoxiong Wang, Long Liu, Yanzhou Yang, Han Xiao, Aojun Zhou, Yafei Wen, Xiaoxin Chen, Shuai Ren, Hongsheng Li
Abstract: The emergence and growing popularity of multimodal large language models (MLLMs) have significant potential to enhance various aspects of daily life, from improving communication to facilitating learning and problem-solving. Mobile phones, as essential daily companions, represent the most effective and accessible deployment platform for MLLMs, enabling seamless integration into everyday tasks. However, deploying MLLMs on mobile phones presents challenges due to limitations in memory size and computational capability, making it difficult to achieve smooth and real-time processing without extensive optimization. In this paper, we present BlueLM-V-3B, an algorithm and system co-design approach specifically tailored for the efficient deployment of MLLMs on mobile platforms. To be specific, we redesign the dynamic resolution scheme adopted by mainstream MLLMs and implement system optimization for hardware-aware deployment to optimize model inference on mobile phones. BlueLM-V-3B boasts the following key highlights: (1) Small Size: BlueLM-V-3B features a language model with 2.7B parameters and a vision encoder with 400M parameters. (2) Fast Speed: BlueLM-V-3B achieves a generation speed of 24.4 token/s on the MediaTek Dimensity 9300 processor with 4-bit LLM weight quantization. (3) Strong Performance: BlueLM-V-3B has attained the highest average score of 66.1 on the OpenCompass benchmark among models with $\leq$ 4B parameters and surpassed a series of models with much larger parameter sizes (e.g., MiniCPM-V-2.6, InternVL2-8B).
Authors: Tanjina Sultana Camelia, Faizur Rahman Fahim, Md. Musfique Anwar
Abstract: Nowadays, the rapid diffusion of fake news poses a significant problem, as it can spread misinformation and confusion. This paper aims to develop an advanced machine learning solution for detecting fake news articles. Leveraging a comprehensive dataset of news articles, including 23,502 fake news articles and 21,417 accurate news articles, we implemented and evaluated three machine-learning models. Our dataset, curated from diverse sources, provides rich textual content categorized into title, text, subject, and Date features. These features are essential for training robust classification models to distinguish between fake and authentic news articles. The initial model employed a Long Short-Term Memory (LSTM) network, achieving an accuracy of 94%. The second model improved upon this by incorporating additional regularization techniques and fine-tuning hyperparameters, resulting in a 97% accuracy. The final model combined the strengths of previous architectures with advanced optimization strategies, achieving a peak accuracy of 98%. These results demonstrate the effectiveness of our approach in identifying fake news with high precision. Implementing these models showcases significant advancements in natural language processing and machine learning techniques, contributing valuable tools for combating misinformation. Our work highlights the potential for deploying such models in real-world applications, providing a reliable method for automated fake news detection and enhancing the credibility of news dissemination.
Authors: Shuyang Hou, Haoyue Jiao, Zhangxiao Shen, Jianyuan Liang, Anqi Zhao, Xiaopu Zhang, Jianxun Wang, Huayi Wu
Abstract: With the rapid growth of interdisciplinary demands for geospatial modeling and the rise of large language models (LLMs), geospatial code generation technology has seen significant advancements. However, existing LLMs often face challenges in the geospatial code generation process due to incomplete or unclear user requirements and insufficient knowledge of specific platform syntax rules, leading to the generation of non-executable code, a phenomenon known as "code hallucination." To address this issue, this paper proposes a Chain of Programming (CoP) framework, which decomposes the code generation process into five steps: requirement analysis, algorithm design, code implementation, code debugging, and code annotation. The framework incorporates a shared information pool, knowledge base retrieval, and user feedback mechanisms, forming an end-to-end code generation flow from requirements to code without the need for model fine-tuning. Based on a geospatial problem classification framework and evaluation benchmarks, the CoP strategy significantly improves the logical clarity, syntactical correctness, and executability of the generated code, with improvements ranging from 3.0% to 48.8%. Comparative and ablation experiments further validate the superiority of the CoP strategy over other optimization approaches and confirm the rationality and necessity of its key components. Through case studies on building data visualization and fire data analysis, this paper demonstrates the application and effectiveness of CoP in various geospatial scenarios. The CoP framework offers a systematic, step-by-step approach to LLM-based geospatial code generation tasks, significantly enhancing code generation performance in geospatial tasks and providing valuable insights for code generation in other vertical domains.
Authors: Seyed Ali Farokh
Abstract: This paper presents our submissions to the Iranian division of the Text-dependent Speaker Verification Challenge (TdSV) 2024. TdSV aims to determine if a specific phrase was spoken by a target speaker. We developed two independent subsystems based on pre-trained models: For phrase verification, a phrase classifier rejected incorrect phrases, while for speaker verification, a pre-trained ResNet293 with domain adaptation extracted speaker embeddings for computing cosine similarity scores. In addition, we evaluated Whisper-PMFA, a pre-trained ASR model adapted for speaker verification, and found that, although it outperforms randomly initialized ResNets, it falls short of the performance of pre-trained ResNets, highlighting the importance of large-scale pre-training. The results also demonstrate that achieving competitive performance on TdSV without joint modeling of speaker and text is possible. Our best system achieved a MinDCF of 0.0358 on the evaluation subset and won the challenge.
Authors: Surasakdi Siripong, Apirak Chaiyapan, Thanakorn Phonchai
Abstract: Remote Sensing Visual Question Answering (RSVQA) is a challenging task that involves interpreting complex satellite imagery to answer natural language questions. Traditional approaches often rely on separate visual feature extractors and language processing models, which can be computationally intensive and limited in their ability to handle open-ended questions. In this paper, we propose a novel method that leverages a generative Large Vision-Language Model (LVLM) to streamline the RSVQA process. Our approach consists of a two-step training strategy: domain-adaptive pretraining and prompt-based finetuning. This method enables the LVLM to generate natural language answers by conditioning on both visual and textual inputs, without the need for predefined answer categories. We evaluate our model on the RSVQAxBEN dataset, demonstrating superior performance compared to state-of-the-art baselines. Additionally, a human evaluation study shows that our method produces answers that are more accurate, relevant, and fluent. The results highlight the potential of generative LVLMs in advancing the field of remote sensing analysis.
Authors: Wenjun Hou, Yi Cheng, Kaishuai Xu, Yan Hu, Wenjie Li, Jiang Liu
Abstract: Comprehensively understanding surgical scenes in Surgical Visual Question Answering (Surgical VQA) requires reasoning over multiple objects. Previous approaches address this task using cross-modal fusion strategies to enhance reasoning ability. However, these methods often struggle with limited scene understanding and question comprehension, and some rely on external resources (e.g., pre-extracted object features), which can introduce errors and generalize poorly across diverse surgical environments. To address these challenges, we propose SCAN, a simple yet effective memory-augmented framework that leverages Multimodal LLMs to improve surgical context comprehension via Self-Contained Inquiry. SCAN operates autonomously, generating two types of memory for context augmentation: Direct Memory (DM), which provides multiple candidates (or hints) to the final answer, and Indirect Memory (IM), which consists of self-contained question-hint pairs to capture broader scene context. DM directly assists in answering the question, while IM enhances understanding of the surgical scene beyond the immediate query. Reasoning over these object-aware memories enables the model to accurately interpret images and respond to questions. Extensive experiments on three publicly available Surgical VQA datasets demonstrate that SCAN achieves state-of-the-art performance, offering improved accuracy and robustness across various surgical scenarios.
Authors: Yangxinyu Xie, Xiang Li, Tanwi Mallick, Weijie J. Su, Ruixun Zhang
Abstract: Watermarking language models is essential for distinguishing between human and machine-generated text and thus maintaining the integrity and trustworthiness of digital communication. We present a novel green/red list watermarking approach that partitions the token set into ``green'' and ``red'' lists, subtly increasing the generation probability for green tokens. To correct token distribution bias, our method employs maximal coupling, using a uniform coin flip to decide whether to apply bias correction, with the result embedded as a pseudorandom watermark signal. Theoretical analysis confirms this approach's unbiased nature and robust detection capabilities. Experimental results show that it outperforms prior techniques by preserving text quality while maintaining high detectability, and it demonstrates resilience to targeted modifications aimed at improving text quality. This research provides a promising watermarking solution for language models, balancing effective detection with minimal impact on text quality.
Authors: Harshita Sharma, Valentina Salvatelli, Shaury Srivastav, Kenza Bouzid, Shruthi Bannur, Daniel C. Castro, Maximilian Ilse, Sam Bond-Taylor, Mercy Prasanna Ranjit, Fabian Falck, Fernando P\'erez-Garc\'ia, Anton Schwaighofer, Hannah Richardson, Maria Teodora Wetscherek, Stephanie L. Hyland, Javier Alvarez-Valle
Abstract: There is growing interest in applying AI to radiology report generation, particularly for chest X-rays (CXRs). This paper investigates whether incorporating pixel-level information through segmentation masks can improve fine-grained image interpretation of multimodal large language models (MLLMs) for radiology report generation. We introduce MAIRA-Seg, a segmentation-aware MLLM framework designed to utilize semantic segmentation masks alongside CXRs for generating radiology reports. We train expert segmentation models to obtain mask pseudolabels for radiology-specific structures in CXRs. Subsequently, building on the architectures of MAIRA, a CXR-specialised model for report generation, we integrate a trainable segmentation tokens extractor that leverages these mask pseudolabels, and employ mask-aware prompting to generate draft radiology reports. Our experiments on the publicly available MIMIC-CXR dataset show that MAIRA-Seg outperforms non-segmentation baselines. We also investigate set-of-marks prompting with MAIRA and find that MAIRA-Seg consistently demonstrates comparable or superior performance. The results confirm that using segmentation masks enhances the nuanced reasoning of MLLMs, potentially contributing to better clinical outcomes.
Authors: Navita Goyal, Ivan Stelmakh, Nihar Shah, Hal Daum\'e III
Abstract: A large host of scientific journals and conferences solicit peer reviews from multiple reviewers for the same submission, aiming to gather a broader range of perspectives and mitigate individual biases. In this work, we reflect on the role of diversity in the slate of reviewers assigned to evaluate a submitted paper as a factor in diversifying perspectives and improving the utility of the peer-review process. We propose two measures for assessing review utility: review coverage -- reviews should cover most contents of the paper -- and review redundancy -- reviews should add information not already present in other reviews. We hypothesize that reviews from diverse reviewers will exhibit high coverage and low redundancy. We conduct a causal study of different measures of reviewer diversity on review coverage and redundancy using observational data from a peer-reviewed conference with approximately 5,000 submitted papers. Our study reveals disparate effects of different diversity measures on review coverage and redundancy. Our study finds that assigning a group of reviewers that are topically diverse, have different seniority levels, or have distinct publication networks leads to broader coverage of the paper or review criteria, but we find no evidence of an increase in coverage for reviewer slates with reviewers from diverse organizations or geographical locations. Reviewers from different organizations, seniority levels, topics, or publications networks (all except geographical diversity) lead to a decrease in redundancy in reviews. Furthermore, publication network-based diversity alone also helps bring in varying perspectives (that is, low redundancy), even within specific review criteria. Our study adopts a group decision-making perspective for reviewer assignments in peer review and suggests dimensions of diversity that can help guide the reviewer assignment process.
Authors: Omar Naim, Guilhem Fouilh\'e, Nicholas Asher
Abstract: In context learning (ICL) is an attractive method of solving a wide range of problems. Inspired by Garg et al. (2022), we look closely at ICL in a variety of train and test settings for several transformer models of different sizes trained from scratch. Our study complements prior work by pointing out several systematic failures of these models to generalize to data not in the training distribution, thereby showing some limitations of ICL. We find that models adopt a strategy for this task that is very different from standard solutions.
Authors: Xinyan Guan, Yanjiang Liu, Xinyu Lu, Boxi Cao, Ben He, Xianpei Han, Le Sun, Jie Lou, Bowen Yu, Yaojie Lu, Hongyu Lin
Abstract: The evolution of machine learning has increasingly prioritized the development of powerful models and more scalable supervision signals. However, the emergence of foundation models presents significant challenges in providing effective supervision signals necessary for further enhancing their capabilities. Consequently, there is an urgent need to explore novel supervision signals and technical approaches. In this paper, we propose verifier engineering, a novel post-training paradigm specifically designed for the era of foundation models. The core of verifier engineering involves leveraging a suite of automated verifiers to perform verification tasks and deliver meaningful feedback to foundation models. We systematically categorize the verifier engineering process into three essential stages: search, verify, and feedback, and provide a comprehensive review of state-of-the-art research developments within each stage. We believe that verifier engineering constitutes a fundamental pathway toward achieving Artificial General Intelligence.
Authors: Longju Bai, Angana Borah, Oana Ignat, Rada Mihalcea
Abstract: Large Multimodal Models (LMMs) exhibit impressive performance across various multimodal tasks. However, their effectiveness in cross-cultural contexts remains limited due to the predominantly Western-centric nature of most data and models. Conversely, multi-agent models have shown significant capability in solving complex tasks. Our study evaluates the collective performance of LMMs in a multi-agent interaction setting for the novel task of cultural image captioning. Our contributions are as follows: (1) We introduce MosAIC, a Multi-Agent framework to enhance cross-cultural Image Captioning using LMMs with distinct cultural personas; (2) We provide a dataset of culturally enriched image captions in English for images from China, India, and Romania across three datasets: GeoDE, GD-VCR, CVQA; (3) We propose a culture-adaptable metric for evaluating cultural information within image captions; and (4) We show that the multi-agent interaction outperforms single-agent models across different metrics, and offer valuable insights for future research. Our dataset and models can be accessed at https://github.com/MichiganNLP/MosAIC.
Authors: Mathew Jacob, Erik Lindgren, Matei Zaharia, Michael Carbin, Omar Khattab, Andrew Drozdov
Abstract: Rerankers, typically cross-encoders, are often used to re-score the documents retrieved by cheaper initial IR systems. This is because, though expensive, rerankers are assumed to be more effective. We challenge this assumption by measuring reranker performance for full retrieval, not just re-scoring first-stage retrieval. Our experiments reveal a surprising trend: the best existing rerankers provide diminishing returns when scoring progressively more documents and actually degrade quality beyond a certain limit. In fact, in this setting, rerankers can frequently assign high scores to documents with no lexical or semantic overlap with the query. We hope that our findings will spur future research to improve reranking.
Authors: Marek Wydmuch, {\L}ukasz Borchmann, Filip Grali\'nski
Abstract: Though large language models (LLMs) have demonstrated exceptional performance across numerous problems, their application to predictive tasks in relational databases remains largely unexplored. In this work, we address the notion that LLMs cannot yield satisfactory results on relational databases due to their interconnected tables, complex relationships, and heterogeneous data types. Using the recently introduced RelBench benchmark, we demonstrate that even a straightforward application of LLMs achieves competitive performance on these tasks. These findings establish LLMs as a promising new baseline for ML on relational databases and encourage further research in this direction.
Authors: Priyanka Singh, Vladislav D. Mosin, Ivan P. Yamshchikov
Abstract: Working within specific NLP subdomains presents significant challenges, primarily due to a persistent deficit of data. Stringent privacy concerns and limited data accessibility often drive this shortage. Additionally, the medical domain demands high accuracy, where even marginal improvements in model performance can have profound impacts. In this study, we investigate the potential of vocabulary transfer to enhance model performance in biomedical NLP tasks. Specifically, we focus on vocabulary extension, a technique that involves expanding the target vocabulary to incorporate domain-specific biomedical terms. Our findings demonstrate that vocabulary extension, leads to measurable improvements in both downstream model performance and inference time.
Authors: Christoph Leiter, Piyawat Lertvittayakumjorn, Marina Fomicheva, Wei Zhao, Yang Gao, Steffen Eger
Abstract: Unlike classical lexical overlap metrics such as BLEU, most current evaluation metrics for machine translation (for example, COMET or BERTScore) are based on black-box large language models. They often achieve strong correlations with human judgments, but recent research indicates that the lower-quality classical metrics remain dominant, one of the potential reasons being that their decision processes are more transparent. To foster more widespread acceptance of novel high-quality metrics, explainability thus becomes crucial. In this concept paper, we identify key properties as well as key goals of explainable machine translation metrics and provide a comprehensive synthesis of recent techniques, relating them to our established goals and properties. In this context, we also discuss the latest state-of-the-art approaches to explainable metrics based on generative models such as ChatGPT and GPT4. Finally, we contribute a vision of next-generation approaches, including natural language explanations. We hope that our work can help catalyze and guide future research on explainable evaluation metrics and, mediately, also contribute to better and more transparent machine translation systems.
Authors: Qiao Jin, Zifeng Wang, Charalampos S. Floudas, Fangyuan Chen, Changlin Gong, Dara Bracken-Clarke, Elisabetta Xue, Yifan Yang, Jimeng Sun, Zhiyong Lu
Abstract: Patient recruitment is challenging for clinical trials. We introduce TrialGPT, an end-to-end framework for zero-shot patient-to-trial matching with large language models. TrialGPT comprises three modules: it first performs large-scale filtering to retrieve candidate trials (TrialGPT-Retrieval); then predicts criterion-level patient eligibility (TrialGPT-Matching); and finally generates trial-level scores (TrialGPT-Ranking). We evaluate TrialGPT on three cohorts of 183 synthetic patients with over 75,000 trial annotations. TrialGPT-Retrieval can recall over 90% of relevant trials using less than 6% of the initial collection. Manual evaluations on 1,015 patient-criterion pairs show that TrialGPT-Matching achieves an accuracy of 87.3% with faithful explanations, close to the expert performance. The TrialGPT-Ranking scores are highly correlated with human judgments and outperform the best-competing models by 43.8% in ranking and excluding trials. Furthermore, our user study reveals that TrialGPT can reduce the screening time by 42.6% in patient recruitment. Overall, these results have demonstrated promising opportunities for patient-to-trial matching with TrialGPT.
Authors: Ningyu Zhang, Yunzhi Yao, Bozhong Tian, Peng Wang, Shumin Deng, Mengru Wang, Zekun Xi, Shengyu Mao, Jintian Zhang, Yuansheng Ni, Siyuan Cheng, Ziwen Xu, Xin Xu, Jia-Chen Gu, Yong Jiang, Pengjun Xie, Fei Huang, Lei Liang, Zhiqiang Zhang, Xiaowei Zhu, Jun Zhou, Huajun Chen
Abstract: Large Language Models (LLMs) have shown extraordinary capabilities in understanding and generating text that closely mirrors human communication. However, a primary limitation lies in the significant computational demands during training, arising from their extensive parameterization. This challenge is further intensified by the dynamic nature of the world, necessitating frequent updates to LLMs to correct outdated information or integrate new knowledge, thereby ensuring their continued relevance. Note that many applications demand continual model adjustments post-training to address deficiencies or undesirable behaviors. There is an increasing interest in efficient, lightweight methods for on-the-fly model modifications. To this end, recent years have seen a burgeoning in the techniques of knowledge editing for LLMs, which aim to efficiently modify LLMs' behaviors within specific domains while preserving overall performance across various inputs. In this paper, we first define the knowledge editing problem and then provide a comprehensive review of cutting-edge approaches. Drawing inspiration from educational and cognitive research theories, we propose a unified categorization criterion that classifies knowledge editing methods into three groups: resorting to external knowledge, merging knowledge into the model, and editing intrinsic knowledge. Furthermore, we introduce a new benchmark, KnowEdit, for a comprehensive empirical evaluation of representative knowledge editing approaches. Additionally, we provide an in-depth analysis of knowledge location, which can give a deeper understanding of the knowledge structures inherent within LLMs. Finally, we discuss several potential applications of knowledge editing, outlining its broad and impactful implications.
Authors: Zening Lin, Jiapeng Wang, Teng Li, Wenhui Liao, Dayi Huang, Longfei Xiong, Lianwen Jin
Abstract: Document pair extraction aims to identify key and value entities as well as their relationships from visually-rich documents. Most existing methods divide it into two separate tasks: semantic entity recognition (SER) and relation extraction (RE). However, simply concatenating SER and RE serially can lead to severe error propagation, and it fails to handle cases like multi-line entities in real scenarios. To address these issues, this paper introduces a novel framework, PEneo (Pair Extraction new decoder option), which performs document pair extraction in a unified pipeline, incorporating three concurrent sub-tasks: line extraction, line grouping, and entity linking. This approach alleviates the error accumulation problem and can handle the case of multi-line entities. Furthermore, to better evaluate the model's performance and to facilitate future research on pair extraction, we introduce RFUND, a re-annotated version of the commonly used FUNSD and XFUND datasets, to make them more accurate and cover realistic situations. Experiments on various benchmarks demonstrate PEneo's superiority over previous pipelines, boosting the performance by a large margin (e.g., 19.89%-22.91% F1 score on RFUND-EN) when combined with various backbones like LiLT and LayoutLMv3, showing its effectiveness and generality. Codes and the new annotations are available at https://github.com/ZeningLin/PEneo.
Authors: Dan Zhang, Ziniu Hu, Sining Zhoubian, Zhengxiao Du, Kaiyu Yang, Zihan Wang, Yisong Yue, Yuxiao Dong, Jie Tang
Abstract: Large Language Models (LLMs) have shown promise in assisting scientific discovery. However, such applications are currently limited by LLMs' deficiencies in understanding intricate scientific concepts, deriving symbolic equations, and solving advanced numerical calculations. To bridge these gaps, we introduce SciInstruct, a suite of scientific instructions for training scientific language models capable of college-level scientific reasoning. Central to our approach is a novel self-reflective instruction annotation framework to address the data scarcity challenge in the science domain. This framework leverages existing LLMs to generate step-by-step reasoning for unlabelled scientific questions, followed by a process of self-reflective critic-and-revise. Applying this framework, we curated a diverse and high-quality dataset encompassing physics, chemistry, math, and formal proofs. We analyze the curated SciInstruct from multiple interesting perspectives (e.g., domain, scale, source, question type, answer length, etc.). To verify the effectiveness of SciInstruct, we fine-tuned different language models with SciInstruct, i.e., ChatGLM3 (6B and 32B), Llama3-8B-Instruct, and Mistral-7B: MetaMath, enhancing their scientific and mathematical reasoning capabilities, without sacrificing the language understanding capabilities of the base model. We release all codes and SciInstruct at https://github.com/THUDM/SciGLM.
Authors: Pittawat Taveekitworachai, Febri Abdullah, Ruck Thawonmas
Abstract: This paper presents null-shot prompting. Null-shot prompting exploits hallucination in large language models (LLMs) by instructing LLMs to utilize information from the "Examples" section that never exists within the provided context to perform a task. While reducing hallucination is crucial and non-negligible for daily and critical uses of LLMs, we propose that in the current landscape in which these LLMs still hallucinate, it is possible, in fact, to exploit hallucination to increase performance in performing tasks compared to standard zero-shot prompting. Experiments with eight LLMs show improvements in performance across the majority of eight datasets, including reading comprehension, arithmetic reasoning, and closed-book question answering. The observed inconsistency in increased relative performance across the LLMs also potentially indicates a different degree of inherent hallucination in each model. These differences show that it is possible to utilize null-shot prompting as a way to detect degrees of hallucination in LLMs using existing benchmarking datasets. We also perform ablation studies, including experimenting with a modified version of null-shot prompting that incorporates ideas from zero-shot chain-of-thought prompting, which shows different trends of results.
Authors: Xianzhen Luo, Qingfu Zhu, Zhiming Zhang, Libo Qin, Xuanyu Zhang, Qing Yang, Dongliang Xu, Wanxiang Che
Abstract: Program of Thoughts (PoT) is an approach characterized by its executable intermediate steps, which ensure the accuracy of the logical calculations in the reasoning process. Currently, PoT primarily uses Python. However, relying solely on a single language may result in suboptimal solutions and overlook the potential benefits of other programming languages. In this paper, we conduct comprehensive experiments on the programming languages used in PoT and find that no single language consistently delivers optimal performance across all tasks and models. The effectiveness of each language varies depending on the specific scenarios. Inspired by this, we propose a task and model agnostic approach called MultiPoT, which harnesses strength and diversity from various languages. Experimental results reveal that it significantly outperforms Python Self-Consistency. Furthermore, it achieves comparable or superior performance compared to the best monolingual PoT in almost all tasks across all models. In particular, MultiPoT achieves more than 4.6% improvement on average on ChatGPT (gpt-3.5-turbo-0701).
Authors: Zhiyuan Wang, Jinhao Duan, Chenxi Yuan, Qingyu Chen, Tianlong Chen, Yue Zhang, Ren Wang, Xiaoshuang Shi, Kaidi Xu
Abstract: Uncertainty estimation is crucial for the reliability of safety-critical human and artificial intelligence (AI) interaction systems, particularly in the domain of healthcare engineering. However, a robust and general uncertainty measure for free-form answers has not been well-established in open-ended medical question-answering (QA) tasks, where generative inequality introduces a large number of irrelevant words and sequences within the generated set for uncertainty quantification (UQ), which can lead to biases. This paper introduces Word-Sequence Entropy (WSE), a method that calibrates uncertainty at both the word and sequence levels, considering semantic relevance. WSE quantifies uncertainty in a way that is more closely aligned with the reliability of LLMs during uncertainty quantification (UQ). We compare WSE with six baseline methods on five free-form medical QA datasets, utilizing seven popular large language models (LLMs). Experimental results demonstrate that WSE exhibits superior performance in UQ under two standard criteria for correctness evaluation. Additionally, in terms of real-world medical QA applications, the performance of LLMs is significantly enhanced (e.g., a 6.36% improvement in model accuracy on the COVID-QA dataset) by employing responses with lower uncertainty that are identified by WSE as final answers, without any additional task-specific fine-tuning or architectural modifications.
Authors: Somnath Banerjee, Sayan Layek, Rima Hazra, Animesh Mukherjee
Abstract: In this study, we tackle a growing concern around the safety and ethical use of large language models (LLMs). Despite their potential, these models can be tricked into producing harmful or unethical content through various sophisticated methods, including 'jailbreaking' techniques and targeted manipulation. Our work zeroes in on a specific issue: to what extent LLMs can be led astray by asking them to generate responses that are instruction-centric such as a pseudocode, a program or a software snippet as opposed to vanilla text. To investigate this question, we introduce TechHazardQA, a dataset containing complex queries which should be answered in both text and instruction-centric formats (e.g., pseudocodes), aimed at identifying triggers for unethical responses. We query a series of LLMs -- Llama-2-13b, Llama-2-7b, Mistral-V2 and Mistral 8X7B -- and ask them to generate both text and instruction-centric responses. For evaluation we report the harmfulness score metric as well as judgements from GPT-4 and humans. Overall, we observe that asking LLMs to produce instruction-centric responses enhances the unethical response generation by ~2-38% across the models. As an additional objective, we investigate the impact of model editing using the ROME technique, which further increases the propensity for generating undesirable content. In particular, asking edited LLMs to generate instruction-centric responses further increases the unethical response generation by ~3-16% across the different models.
Authors: Yuan Ge, Yilun Liu, Chi Hu, Weibin Meng, Shimin Tao, Xiaofeng Zhao, Hongxia Ma, Li Zhang, Boxing Chen, Hao Yang, Bei Li, Tong Xiao, Jingbo Zhu
Abstract: With contributions from the open-source community, a vast amount of instruction tuning (IT) data has emerged. Given the significant resource allocation required for training and evaluating models, it is advantageous to have an efficient method for selecting high-quality IT data. However, existing methods for instruction data selection have limitations such as relying on fragile external APIs, being affected by biases in GPT models, or reducing the diversity of the selected instruction dataset. In this paper, we propose an industrial-friendly, expert-aligned and diversity-preserved instruction data selection method: Clustering and Ranking (CaR). CaR employs a two-step process: first, it ranks instruction pairs using a high-accuracy (84.25%) scoring model aligned with expert preferences; second, it preserves dataset diversity through clustering. In our experiment, CaR efficiently selected a mere 1.96% of Alpaca's IT data, yet the resulting AlpaCaR model surpassed Alpaca's performance by an average of 32.1% in GPT-4 evaluations. Moreover, we find that data selecting is a consistent paradigm whether the pre-trained model is more capable or the model parameters scaling up. Our approach employs compact models with 550M parameters and incurs just 11.2% of the financial outlay of current methods, enhancing its industrial deployability.
Authors: Ming Dong, Kang Xue, Bolong Zheng, Tingting He
Abstract: Fine-tuning all parameters of Large Language Models (LLMs) is computationally expensive. Parameter-Efficient Fine-Tuning (PEFT) methods address this by selectively fine-tuning specific parameters. Most of the parameter efficient fine-tuning (PEFT) methods center on selecting or introducing a set of parameters to be fine-tuned. However, there are few methods that consider the impact of data samples on parameter selecting. Representative data driven methods include FISH Mask based method, which randomly selects a portion of data samples as a basis when selecting parameters. However, this random data sample selection method cannot select optimal parameters for unstable data distribution. In this work, we introduce a data-centric approach and propose the Iterative Range Decreasing (IRD) algorithm to optimize the sample-parameter pair selection in FISH Mask. IRD iteratively refines the selection by identifying subsets of samples and parameters exhibiting higher Fisher information. We demonstrate the effectiveness and rationality of proposed strategy by conducting experiments on GLUE benchmark. Experimental results show our strategy optimizes the parameter selection and achieves preferable performance over some typical baseline methods.
Authors: Bo-Ru Lu, Nikita Haduong, Chien-Yu Lin, Hao Cheng, Noah A. Smith, Mari Ostendorf
Abstract: Transformer-based NLP models are powerful but have high computational costs that limit deployment. Finetuned encoder-decoder models are popular in specialized domains and can outperform larger more generalized decoder-only models, such as GPT-4. We introduce a new configuration for encoder-decoder models that improves efficiency on structured output and decomposable tasks where multiple outputs are required for a single shared input. Our method, prompt-in-decoder (PiD), encodes the input once and decodes the output in parallel, boosting both training and inference efficiency by avoiding duplicate input encoding and increasing the operational intensity (ratio of numbers of arithmetic operation to memory access) of decoding process by sharing the input key-value cache. We achieve computation reduction that roughly scales with the number of subtasks, gaining up to 4.6x speed-up over state-of-the-art models for dialogue state tracking, summarization, and question-answering tasks, with comparable or better performance.
Authors: Zekun Wu, Sahan Bulathwela, Maria Perez-Ortiz, Adriano Soares Koshiyama
Abstract: Stereotype detection is a challenging and subjective task, as certain statements, such as "Black people like to play basketball," may not appear overtly toxic but still reinforce racial stereotypes. With the increasing prevalence of large language models (LLMs) in human-facing artificial intelligence (AI) applications, detecting these types of biases is essential. However, LLMs risk perpetuating and amplifying stereotypical outputs derived from their training data. A reliable stereotype detector is crucial for benchmarking bias, monitoring model input and output, filtering training data, and ensuring fairer model behavior in downstream applications. This paper introduces the Multi-Grain Stereotype (MGS) dataset, consisting of 51,867 instances across gender, race, profession, religion, and other stereotypes, curated from multiple existing datasets. We evaluate various machine learning approaches to establish baselines and fine-tune language models of different architectures and sizes, presenting a suite of stereotype multiclass classifiers trained on the MGS dataset. Given the subjectivity of stereotypes, explainability is essential to align model learning with human understanding of stereotypes. We employ explainable AI (XAI) tools, including SHAP, LIME, and BertViz, to assess whether the model's learned patterns align with human intuitions about stereotypes.Additionally, we develop stereotype elicitation prompts and benchmark the presence of stereotypes in text generation tasks using popular LLMs, employing the best-performing stereotype classifiers.
Authors: Lydia Uhler, Verena Jordan, J\"urgen Buder, Markus Huff, Frank Papenmeier
Abstract: Generative artificial intelligences, particularly large language models (LLMs), play an increasingly prominent role in human decision-making contexts, necessitating transparency about their capabilities. While prior studies have shown addition biases in humans (Adams et al., 2021) and OpenAI's GPT-3 (Winter et al., 2023), this study extends the research by comparing human and GPT-4 problem-solving across both spatial and linguistic tasks, with variations in solution efficiency and valence of task instruction. Four preregistered experiments with 588 participants from the U.S. and 680 GPT-4 iterations revealed a stronger tendency towards additive transformations in GPT-4 than in humans. Human participants were less likely to use additive strategies when subtraction was relatively more efficient than when addition and subtraction were equally efficient. GPT-4 exhibited the opposite behavior, with a strong addition bias when subtraction was more efficient. In terms of valence of task instruction, GPT-4's use of additive strategies increased when instructed to "improve" (positive) rather than "edit" (neutral). These findings demonstrate that biases in human problem-solving are amplified in GPT-4, and that LLM behavior differs from human efficiency-based strategies. This highlights the limitations of LLMs and the need for caution when using them in real-world applications.
Authors: D. Subhalingam, Keshav Kolluru, Mausam, Saurabh Singal
Abstract: In the e-commerce domain, the accurate extraction of attribute-value pairs (e.g., Brand: Apple) from product titles and user search queries is crucial for enhancing search and recommendation systems. A major challenge with neural models for this task is the lack of high-quality training data, as the annotations for attribute-value pairs in the available datasets are often incomplete. To address this, we introduce GenToC, a model designed for training directly with partially-labeled data, eliminating the necessity for a fully annotated dataset. GenToC employs a marker-augmented generative model to identify potential attributes, followed by a token classification model that determines the associated values for each attribute. GenToC outperforms existing state-of-the-art models, exhibiting upto 56.3% increase in the number of accurate extractions. Furthermore, we utilize GenToC to regenerate the training dataset to expand attribute-value annotations. This bootstrapping substantially improves the data quality for training other standard NER models, which are typically faster but less capable in handling partially-labeled data, enabling them to achieve comparable performance to GenToC. Our results demonstrate GenToC's unique ability to learn from a limited set of partially-labeled data and improve the training of more efficient models, advancing the automated extraction of attribute-value pairs. Finally, our model has been successfully integrated into IndiaMART, India's largest B2B e-commerce platform, achieving a significant increase of 20.2% in the number of correctly identified attribute-value pairs over the existing deployed system while achieving a high precision of 89.5%.
Authors: Zican Dong, Junyi Li, Xin Men, Wayne Xin Zhao, Bingbing Wang, Zhen Tian, Weipeng Chen, Ji-Rong Wen
Abstract: Transformer-based large language models (LLMs) typically have a limited context window, resulting in significant performance degradation when processing text beyond the length of the context window. Extensive studies have been proposed to extend the context window and achieve length extrapolation of LLMs, but there is still a lack of in-depth interpretation of these approaches. In this study, we explore the positional information within and beyond the context window for deciphering the underlying mechanism of LLMs. By using a mean-based decomposition method, we disentangle positional vectors from hidden states of LLMs and analyze their formation and effect on attention. Furthermore, when texts exceed the context window, we analyze the change of positional vectors in two settings, i.e., direct extrapolation and context window extension. Based on our findings, we design two training-free context window extension methods, positional vector replacement and attention window extension. Experimental results show that our methods can effectively extend the context window length.
Authors: Felix B Mueller, Rebekka G\"orge, Anna K Bernzen, Janna C Pirk, Maximilian Poretschkin
Abstract: Memorization in large language models (LLMs) is a growing concern. LLMs have been shown to easily reproduce parts of their training data, including copyrighted work. This is an important problem to solve, as it may violate existing copyright laws as well as the European AI Act. In this work, we propose a systematic analysis to quantify the extent of potential copyright infringements in LLMs using European law as an example. Unlike previous work, we evaluate instruction-finetuned models in a realistic end-user scenario. Our analysis builds on a proposed threshold of 160 characters, which we borrow from the German Copyright Service Provider Act and a fuzzy text matching algorithm to identify potentially copyright-infringing textual reproductions. The specificity of countermeasures against copyright infringement is analyzed by comparing model behavior on copyrighted and public domain data. We investigate what behaviors models show instead of producing protected text (such as refusal or hallucination) and provide a first legal assessment of these behaviors. We find that there are huge differences in copyright compliance, specificity, and appropriate refusal among popular LLMs. Alpaca, GPT 4, GPT 3.5, and Luminous perform best in our comparison, with OpenGPT-X, Alpaca, and Luminous producing a particularly low absolute number of potential copyright violations. Code can be found at https://github.com/felixbmuller/llms-memorization-copyright.
URLs: https://github.com/felixbmuller/llms-memorization-copyright.
Authors: Dan Zhang, Sining Zhoubian, Ziniu Hu, Yisong Yue, Yuxiao Dong, Jie Tang
Abstract: Recent methodologies in LLM self-training mostly rely on LLM generating responses and filtering those with correct output answers as training data. This approach often yields a low-quality fine-tuning training set (e.g., incorrect plans or intermediate reasoning). In this paper, we develop a reinforced self-training approach, called ReST-MCTS*, based on integrating process reward guidance with tree search MCTS* for collecting higher-quality reasoning traces as well as per-step value to train policy and reward models. ReST-MCTS* circumvents the per-step manual annotation typically used to train process rewards by tree-search-based reinforcement learning: Given oracle final correct answers, ReST-MCTS* is able to infer the correct process rewards by estimating the probability this step can help lead to the correct answer. These inferred rewards serve dual purposes: they act as value targets for further refining the process reward model and also facilitate the selection of high-quality traces for policy model self-training. We first show that the tree-search policy in ReST-MCTS* achieves higher accuracy compared with prior LLM reasoning baselines such as Best-of-N and Tree-of-Thought, within the same search budget. We then show that by using traces searched by this tree-search policy as training data, we can continuously enhance the three language models for multiple iterations, and outperform other self-training algorithms such as ReST$^\text{EM}$ and Self-Rewarding LM. We release all code at https://github.com/THUDM/ReST-MCTS.
Authors: Zichun Yu, Spandan Das, Chenyan Xiong
Abstract: Pretraining data selection has the potential to improve language model pretraining efficiency by utilizing higher-quality data from massive web data corpora. Current data selection methods, which rely on either hand-crafted rules or larger reference models, are conducted statically and do not capture the evolving data preferences during pretraining. In this paper, we introduce model-aware data selection with data influence models (MATES), where a data influence model continuously adapts to the evolving data preferences of the pretraining model and then selects the data most effective for the current pretraining progress. Specifically, we collect oracle data influence by locally probing the pretraining model and fine-tune a small data influence model to approximate it accurately. The data influence model then predicts data influence over the whole pretraining corpus and selects the most influential data for the next pretraining stage. Experiments of pretraining 410M and 1B models on the C4 dataset demonstrate that MATES significantly outperforms random data selection on extensive downstream tasks. It doubles the gains achieved by the state-of-the-art data selection approach that leverages larger reference models and reduces the total FLOPs required to reach certain performances by half. Further analyses validate the effectiveness of the locally probed oracle data influence and the approximation with data influence models. Our code is open-sourced at https://github.com/cxcscmu/MATES.
Authors: Julen Etxaniz, Gorka Azkune, Aitor Soroa, Oier Lopez de Lacalle, Mikel Artetxe
Abstract: Large Language Models (LLMs) exhibit extensive knowledge about the world, but most evaluations have been limited to global or anglocentric subjects. This raises the question of how well these models perform on topics relevant to other cultures, whose presence on the web is not that prominent. To address this gap, we introduce BertaQA, a multiple-choice trivia dataset that is parallel in English and Basque. The dataset consists of a local subset with questions pertinent to the Basque culture, and a global subset with questions of broader interest. We find that state-of-the-art LLMs struggle with local cultural knowledge, even as they excel on global topics. However, we show that continued pre-training in Basque significantly improves the models' performance on Basque culture, even when queried in English. To our knowledge, this is the first solid evidence of knowledge transfer from a low-resource to a high-resource language. Our analysis sheds light on the complex interplay between language and knowledge, and reveals that some prior findings do not fully hold when reassessed on local topics. Our dataset and evaluation code are available under open licenses at https://github.com/juletx/BertaQA.
Authors: Marcos Piau, Roberto Lotufo, Rodrigo Nogueira
Abstract: Despite advancements in Natural Language Processing (NLP) and the growing availability of pretrained models, the English language remains the primary focus of model development. Continued pretraining on language-specific corpora provides a practical solution for adapting models to other languages. However, the impact of different pretraining settings on downstream tasks remains underexplored. This work introduces $\texttt{ptt5-v2}$, investigating the continued pretraining of T5 models for Portuguese. We first develop a baseline set of settings and pretrain models with sizes up to 3B parameters. Finetuning on three Portuguese downstream tasks (assin2 STS, assin2 RTE, and TweetSentBR) yields SOTA results on the latter two. We then explore the effects of different pretraining configurations, including pretraining data quality, optimization strategies, and multi-epoch pretraining. Perhaps surprisingly, their impact remains subtle compared to our baseline. We release $\texttt{ptt5-v2}$ pretrained checkpoints and their MonoT5-based finetuned $\texttt{MonoPTT5}$ rerankers on HuggingFace in their respective collections at \url{https://huggingface.co/unicamp-dl}.
Authors: Jessica Zhu, Iain Cruickshank, Michel Cukier
Abstract: News will have biases so long as people have opinions. It is increasingly important for informed citizens to be able to identify bias as social media becomes the primary entry point for news and partisan differences increase. If people know the biases of the news they are consuming, they will be able to take action to avoid polarizing echo chambers. In this paper, we explore an often overlooked aspect of bias detection in documents: the semantic structure of news articles. We present DocNet, a novel, inductive, and low-resource document embedding and bias detection model that outperforms large language models. We also demonstrate that the semantic structure of news articles from opposing partisan sides, as represented in document-level graph embeddings, have significant similarities. These results can be used to advance bias detection in low-resource environments. Our code, data, and the corresponding datasheet are made available at: https://anonymous.4open.science/r/DocNet/.
Authors: Amit Das, Zheng Zhang, Najib Hasan, Souvika Sarkar, Fatemeh Jamshidi, Tathagata Bhattacharya, Mostafa Rahgouy, Nilanjana Raychawdhary, Dongji Feng, Vinija Jain, Aman Chadha, Mary Sandage, Lauramarie Pope, Gerry Dozier, Cheryl Seals
Abstract: Data annotation, the practice of assigning descriptive labels to raw data, is pivotal in optimizing the performance of machine learning models. However, it is a resource-intensive process susceptible to biases introduced by annotators. The emergence of sophisticated Large Language Models (LLMs) presents a unique opportunity to modernize and streamline this complex procedure. While existing research extensively evaluates the efficacy of LLMs, as annotators, this paper delves into the biases present in LLMs when annotating hate speech data. Our research contributes to understanding biases in four key categories: gender, race, religion, and disability with four LLMs: GPT-3.5, GPT-4o, Llama-3.1 and Gemma-2. Specifically targeting highly vulnerable groups within these categories, we analyze annotator biases. Furthermore, we conduct a comprehensive examination of potential factors contributing to these biases by scrutinizing the annotated data. We introduce our custom hate speech detection dataset, HateBiasNet, to conduct this research. Additionally, we perform the same experiments on the ETHOS (Mollas et al. 2022) dataset also for comparative analysis. This paper serves as a crucial resource, guiding researchers and practitioners in harnessing the potential of LLMs for data annotation, thereby fostering advancements in this critical field.
Authors: Ukyo Honda, Tatsushi Oka, Peinan Zhang, Masato Mita
Abstract: Recent models for natural language understanding are inclined to exploit simple patterns in datasets, commonly known as shortcuts. These shortcuts hinge on spurious correlations between labels and latent features existing in the training data. At inference time, shortcut-dependent models are likely to generate erroneous predictions under distribution shifts, particularly when some latent features are no longer correlated with the labels. To avoid this, previous studies have trained models to eliminate the reliance on shortcuts. In this study, we explore a different direction: pessimistically aggregating the predictions of a mixture-of-experts, assuming each expert captures relatively different latent features. The experimental results demonstrate that our post-hoc control over the experts significantly enhances the model's robustness to the distribution shift in shortcuts. Besides, we show that our approach has some practical advantages. We also analyze our model and provide results to support the assumption.
Authors: Sumit Kumar Dam, Choong Seon Hong, Yu Qiao, Chaoning Zhang
Abstract: The past few decades have witnessed an upsurge in data, forming the foundation for data-hungry, learning-based AI technology. Conversational agents, often referred to as AI chatbots, rely heavily on such data to train large language models (LLMs) and generate new content (knowledge) in response to user prompts. With the advent of OpenAI's ChatGPT, LLM-based chatbots have set new standards in the AI community. This paper presents a complete survey of the evolution and deployment of LLM-based chatbots in various sectors. We first summarize the development of foundational chatbots, followed by the evolution of LLMs, and then provide an overview of LLM-based chatbots currently in use and those in the development phase. Recognizing AI chatbots as tools for generating new knowledge, we explore their diverse applications across various industries. We then discuss the open challenges, considering how the data used to train the LLMs and the misuse of the generated knowledge can cause several issues. Finally, we explore the future outlook to augment their efficiency and reliability in numerous applications. By addressing key milestones and the present-day context of LLM-based chatbots, our survey invites readers to delve deeper into this realm, reflecting on how their next generation will reshape conversational AI.
Authors: Christoph Leiter, Steffen Eger
Abstract: Large language models (LLMs) have revolutionized NLP research. Notably, in-context learning enables their use as evaluation metrics for natural language generation, making them particularly advantageous in low-resource scenarios and time-restricted applications. In this work, we introduce PrExMe, a large-scale Prompt Exploration for Metrics, where we evaluate more than 720 prompt templates for open-source LLM-based metrics on machine translation (MT) and summarization datasets, totalling over 6.6M evaluations. This extensive comparison (1) benchmarks recent open-source LLMs as metrics and (2) explores the stability and variability of different prompting strategies. We discover that, on the one hand, there are scenarios for which prompts are stable. For instance, some LLMs show idiosyncratic preferences and favor to grade generated texts with textual labels while others prefer to return numeric scores. On the other hand, the stability of prompts and model rankings can be susceptible to seemingly innocuous changes. For example, changing the requested output format from "0 to 100" to "-1 to +1" can strongly affect the rankings in our evaluation. Our study contributes to understanding the impact of different prompting approaches on LLM-based metrics for MT and summarization evaluation, highlighting the most stable prompting patterns and potential limitations.
Authors: Zhiyuan Wang, Jinhao Duan, Lu Cheng, Yue Zhang, Qingni Wang, Xiaoshuang Shi, Kaidi Xu, Hengtao Shen, Xiaofeng Zhu
Abstract: Uncertainty quantification (UQ) in natural language generation (NLG) tasks remains an open challenge, exacerbated by the closed-source nature of the latest large language models (LLMs). This study investigates applying conformal prediction (CP), which can transform any heuristic uncertainty notion into rigorous prediction sets, to black-box LLMs in open-ended NLG tasks. We introduce a novel uncertainty measure based on self-consistency theory, and then develop a conformal uncertainty criterion by integrating the uncertainty condition aligned with correctness into the CP algorithm. Empirical evaluations indicate that our uncertainty measure outperforms prior state-of-the-art methods. Furthermore, we achieve strict control over the correctness coverage rate utilizing 7 popular LLMs on 4 free-form NLG datasets, spanning general-purpose and medical scenarios. Additionally, the calibrated prediction sets with small size further highlights the efficiency of our method in providing trustworthy guarantees for practical open-ended NLG applications.
Authors: Orevaoghene Ahia, Sachin Kumar, Hila Gonen, Valentin Hofmann, Tomasz Limisiewicz, Yulia Tsvetkov, Noah A. Smith
Abstract: In multilingual settings, non-Latin scripts and low-resource languages are usually disadvantaged in terms of language models' utility, efficiency, and cost. Specifically, previous studies have reported multiple modeling biases that the current tokenization algorithms introduce to non-Latin script languages, the main one being over-segmentation. In this work, we propose MAGNET; multilingual adaptive gradient-based tokenization to reduce over-segmentation via adaptive gradient-based subword tokenization. MAGNET learns to predict segment boundaries between byte tokens in a sequence via sub-modules within the model, which act as internal boundary predictors (tokenizers). Previous gradient-based tokenization methods aimed for uniform compression across sequences by integrating a single boundary predictor during training and optimizing it end-to-end through stochastic reparameterization alongside the next token prediction objective. However, this approach still results in over-segmentation for non-Latin script languages in multilingual settings. In contrast, MAGNET offers a customizable architecture where byte-level sequences are routed through language-script-specific predictors, each optimized for its respective language script. This modularity enforces equitable segmentation granularity across different language scripts compared to previous methods. Through extensive experiments, we demonstrate that in addition to reducing segmentation disparities, MAGNET also enables faster language modelling and improves downstream utility.
Authors: Branden Butler, Sixing Yu, Arya Mazaheri, Ali Jannesari
Abstract: Inference of Large Language Models (LLMs) across computer clusters has become a focal point of research in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce bottlenecks associated with memory bandwidth, but also increase end-to-end latency per inference run, requiring high speculation acceptance rates to improve performance. Combined with a variable rate of acceptance across tasks, speculative inference techniques can result in reduced performance. Additionally, pipeline-parallel designs require many user requests to maintain maximum utilization. As a remedy, we propose PipeInfer, a pipelined speculative acceleration technique to reduce inter-token latency and improve system utilization for single-request scenarios while also improving tolerance to low speculation acceptance rates and low-bandwidth interconnects. PipeInfer exhibits up to a 2.15$\times$ improvement in generation speed over standard speculative inference. PipeInfer achieves its improvement through Continuous Asynchronous Speculation and Early Inference Cancellation, the former improving latency and generation speed by running single-token inference simultaneously with several speculative runs, while the latter improves speed and latency by skipping the computation of invalidated runs, even in the middle of inference.
Authors: Jiahuan Li, Shujian Huang, Aarron Ching, Xinyu Dai, Jiajun Chen
Abstract: Large language models demonstrate reasonable multilingual abilities, despite predominantly English-centric pretraining. However, the spontaneous multilingual alignment in these models is shown to be weak, leading to unsatisfactory cross-lingual transfer and knowledge sharing. Previous works attempt to address this issue by explicitly injecting multilingual alignment information during or after pretraining. Thus for the early stage in pretraining, the alignment is weak for sharing information or knowledge across languages. In this paper, we propose PreAlign, a framework that establishes multilingual alignment prior to language model pretraining. PreAlign injects multilingual alignment by initializing the model to generate similar representations of aligned words and preserves this alignment using a code-switching strategy during pretraining. Extensive experiments in a synthetic English to English-Clone setting demonstrate that PreAlign significantly outperforms standard multilingual joint training in language modeling, zero-shot cross-lingual transfer, and cross-lingual knowledge application. Further experiments in real-world scenarios further validate PreAlign's effectiveness across various model sizes.
Authors: Sohini Roychowdhury, Marko Krema, Brian Moore, Xingjian Lai, Dike Effedua, Bharat Jethwani
Abstract: Recent trends in Generative AI have emerged towards fine-tuning foundational large language models (LLMs) to create domain-specific LLMs for automation and chatbot-like applications. Specialized applications for analytics-heavy domains such as Financial report generation require specific writing styles that comprise compound and creative sentences with minimized hallucinations. In this work, we explore the self-corrective auto-regressive qualities of LLMs to learn creativity in writing styles with minimal prompting. We propose a novel two-stage fine-tuning (FT) strategy wherein in the first stage public domain financial reports are used to train for writing styles while allowing the LLM to hallucinate. In the second stage the examples of hallucinations are manually corrected and further used to fine-tune the LLM. The finally trained LLM learns to generate specific financial report sections using minimal instructions and tabular data inputs while ensuring low fine-tuning costs. Our proposed two-stage fine-tuning boosts the accuracy of financial questions answering by two-folds while reducing hallucinations by over 50%. Also, the fine-tuned model has lower perplexity, improved ROUGE, TER and BLEU scores, higher creativity and knowledge density with lower uncertainty and cross entropy than base LLMs. Thus, the proposed framework can be generalized to train creativity in LLMs by first allowing them to hallucinate.
Authors: Ruosen Li, Ruochen Li, Barry Wang, Xinya Du
Abstract: To evaluate Large Language Models (LLMs) for question answering (QA), traditional methods typically focus on assessing single-turn responses to given questions. However, this approach doesn't capture the dynamic nature of human-AI interactions, where humans actively seek information through conversation. Recent works in human-computer interaction (HCI) have employed human evaluators to conduct interactions and evaluations, but they are often prohibitively expensive and time-consuming to scale. We introduce an automatic evaluation framework IQA-EVAL to achieve Interactive Question Answering Evaluations, more specifically, we introduce a LLM-based Evaluation Agent (LEA) that can: (1) simulate human behaviors to generate interactions with IQA models; (2) automatically evaluate the generated interactions. Moreover, we propose assigning personas to LEAs to better simulate groups of real human evaluators. We show that: (1) our evaluation framework with GPT-4 (or Claude) as the backbone model achieves a high correlation with human evaluations on the IQA task; (2) assigning personas to LEA to better represent the crowd further significantly improves correlations. Finally, we use our automatic metric to evaluate five recent representative LLMs with over 1000 questions from complex and ambiguous question answering tasks, which comes with a substantial cost of $5k if evaluated by humans.
Authors: Dongyu Gong, Hantao Zhang
Abstract: Recent work on Transformer-based large language models (LLMs) has revealed striking limits in their working memory capacity, similar to what has been found in human behavioral studies. Specifically, these models' performance drops significantly on N-back tasks as N increases. However, there is still a lack of mechanistic interpretability as to why this phenomenon would arise. Inspired by the executive attention theory from behavioral sciences, we hypothesize that the self-attention mechanism within Transformer-based models might be responsible for their working memory capacity limits. To test this hypothesis, we train vanilla decoder-only transformers to perform N-back tasks and find that attention scores gradually aggregate to the N-back positions over training, suggesting that the model masters the task by learning a strategy to pay attention to the relationship between the current position and the N-back position. Critically, we find that the total entropy of the attention score matrix increases as N increases, suggesting that the dispersion of attention scores might be the cause of the capacity limit observed in N-back tasks. Our findings thus offer insights into the shared role of attention in both human and artificial intelligence. Moreover, the limitations of the self-attention mechanism revealed in the current study could inform future efforts to design more powerful model architectures with enhanced working memory capacity and cognitive capabilities.
Authors: Xin Guan, Nathaniel Demchak, Saloni Gupta, Ze Wang, Ediz Ertekin Jr., Adriano Koshiyama, Emre Kazim, Zekun Wu
Abstract: The development of unbiased large language models is widely recognized as crucial, yet existing benchmarks fall short in detecting biases due to limited scope, contamination, and lack of a fairness baseline. SAGED(-Bias) is the first holistic benchmarking pipeline to address these problems. The pipeline encompasses five core stages: scraping materials, assembling benchmarks, generating responses, extracting numeric features, and diagnosing with disparity metrics. SAGED includes metrics for max disparity, such as impact ratio, and bias concentration, such as Max Z-scores. Noticing that assessment tool bias and contextual bias in prompts can distort evaluation, SAGED implements counterfactual branching and baseline calibration for mitigation. For demonstration, we use SAGED on G20 Countries with popular 8b-level models including Gemma2, Llama3.1, Mistral, and Qwen2. With sentiment analysis, we find that while Mistral and Qwen2 show lower max disparity and higher bias concentration than Gemma2 and Llama3.1, all models are notably biased against countries like Russia and (except for Qwen2) China. With further experiments to have models role-playing U.S. (vice-/former-) presidents, we see bias amplifies and shifts in heterogeneous directions. Moreover, we see Qwen2 and Mistral not engage in role-playing, while Llama3.1 and Gemma2 role-play Trump notably more intensively than Biden and Harris, indicating role-playing performance bias in these models.
Authors: Stefan Heimersheim
Abstract: The LayerNorm (LN) layer in GPT-style transformer models has long been a hindrance to mechanistic interpretability. LN is a crucial component required to stabilize the training of large language models, and LN or the similar RMSNorm have been used in practically all large language models based on the transformer architecture. The non-linear nature of the LN layers is a hindrance for mechanistic interpretability as it hinders interpretation of the residual stream, and makes it difficult to decompose the model into circuits. Some researchers have gone so far as to name "reasons interpretability researchers hate layer norm." In this paper we show that it is possible to remove the LN layers from a pre-trained GPT2-small model by fine-tuning on a fraction (500M tokens) of the training data. We demonstrate that this LN-free model achieves similar performance to the original model on the OpenWebText and ThePile datasets (-0.05 cross-entropy loss), and the Hellaswag benchmark (-0.5% accuracy). We provide our implementation at https://github.com/ApolloResearch/gpt2_noLN, and fine-tuned GPT2-small models at https://huggingface.co/apollo-research/gpt2_noLN. Our work not only provides a simplified model for mechanistic interpretability research, but also provides evidence that the LN layers, at inference time, do not play a crucial role in transformer models.
URLs: https://github.com/ApolloResearch/gpt2_noLN,, https://huggingface.co/apollo-research/gpt2_noLN.
Authors: Eric Cullhed
Abstract: This article presents an experiment in fine-tuning a pretrained causal language model (Meta's Llama 3.1 8B Instruct) to assist with restoring missing or illegible characters in ancient Greek inscriptions and documentary papyri. Utilizing a straightforward instruction-based approach and a 95%/5% train/test split, the papyrus restoration model achieved a character error rate (CER) of 14.9%, a top-1 accuracy of 73.5%, and a top-20 accuracy of 86.0% for sequences up to 10 characters. A model was also fine-tuned for geographic attribution, reaching a top-1 accuracy of 66.4% and a top-3 accuracy of 79.9%. In chronological attribution, it demonstrated an average deviation of 21.7 years from the actual terminus post/ante quem, with a median deviation of 0 years. For inscriptions, the restoration model achieved a CER of 20.5%, a top-1 accuracy of 63.7%, and a top-20 accuracy of 83.0% for sequences up to 10 characters. In geographic attribution, it attained a top-1 accuracy of 75.0% and a top-3 accuracy of 83.7%, while in dating, it had an average deviation of 37.1 years and a median deviation of 3 years from the actual date range. Benchmarked against the state-of-the-art model (Ithaca) on a shared test set and on recently edited inscriptions, the instruction-tuned models excelled in text restoration, while also offering the practical advantage of ignoring spaces during reconstruction, which aligns with the scriptio continua of ancient textual artifacts. However, their performance in geographic and chronological attribution was lower than Ithaca's. To evaluate the approach in a more even setup, the instruction model was retrained with an 80%/10%/10% train-validation-test split, and still outperformed Ithaca in text restoration. The results suggest that fine-tuning larger pretrained causal language models using instruction templates for emendations and conjectures to ancient texts holds promise.
Authors: Emanuela Boros, Maud Ehrmann
Abstract: This paper investigates the presence of OCR-sensitive neurons within the Transformer architecture and their influence on named entity recognition (NER) performance on historical documents. By analysing neuron activation patterns in response to clean and noisy text inputs, we identify and then neutralise OCR-sensitive neurons to improve model performance. Based on two open access large language models (Llama2 and Mistral), experiments demonstrate the existence of OCR-sensitive regions and show improvements in NER performance on historical newspapers and classical commentaries, highlighting the potential of targeted neuron modulation to improve models' performance on noisy text.
Authors: Tianjian Li, Haoran Xu, Weiting Tan, Kenton Murray, Daniel Khashabi
Abstract: Data availability across domains often follows a long-tail distribution: a few domains have abundant data, while most face dat . a scarcity. This imbalance poses challenges in training language models uniformly across all domains. In our study, we focus on multilingual settings, where data sizes vary significantly between high- and low-resource languages. Common strategies to address this include upsampling low-resource languages (Temperature Sampling) or upweighting their loss (Scalarization). Although often considered equivalent, this assumption has not been proven, which motivates our study. Through both theoretical and empirical analysis, we identify the conditions under which these approaches are equivalent and when they diverge. Specifically, we demonstrate that these two methods are equivalent under full gradient descent, but this equivalence breaks down with stochastic gradient descent. Empirically, we observe that Temperature Sampling converges more quickly but is prone to overfitting. We argue that this faster convergence is likely due to the lower variance in gradient estimations, as shown theoretically. Based on these insights, we propose Cooldown, a strategy that reduces sampling temperature during training, accelerating convergence without overfitting to low-resource languages. Our method is competitive with existing data re-weighting and offers computational efficiency.
Authors: Yuelyu Ji, Zhuochun Li, Rui Meng, Daqing He
Abstract: Reranking documents based on their relevance to a given query is a critical task in information retrieval. Traditional reranking methods often lack transparency and rely on proprietary models, hindering reproducibility and interpretability. We propose Reason-to-Rank (R2R), a novel open-source reranking approach that enhances transparency by generating two types of reasoning: direct relevance reasoning, which explains how a document addresses the query, and comparison reasoning, which justifies the relevance of one document over another. We leverage large language models (LLMs) as teacher models to generate these explanations and distill this knowledge into smaller, openly available student models. Our student models are trained to generate meaningful reasoning and rerank documents, achieving competitive performance across multiple datasets, including MSMARCO and BRIGHT. Experiments demonstrate that R2R not only improves reranking accuracy but also provides valuable insights into the decision-making process. By offering a structured and interpretable solution with openly accessible resources, R2R aims to bridge the gap between effectiveness and transparency in information retrieval, fostering reproducibility and further research in the field.
Authors: Xinliang Frederick Zhang, Nick Beauchamp, Lu Wang
Abstract: Reasoning about time and temporal relations is an integral aspect of human cognition, essential for perceiving the world and navigating our experiences. Though large language models (LLMs) have demonstrated impressive performance in many reasoning tasks, temporal reasoning remains challenging due to its intrinsic complexity. In this work, we first study an essential task of temporal reasoning -- temporal graph generation, to unveil LLMs' inherent, global reasoning capabilities. We show that this task presents great challenges even for the most powerful LLMs, such as GPT-3.5/4. We also notice a significant performance gap by small models (<10B) that lag behind LLMs by 50%. Next, we study how to close this gap with a budget constraint, e.g., not using model finetuning. We propose a new prompting technique tailored for temporal reasoning, Narrative-of-Thought (NoT), that first converts the events set to a Python class, then prompts a small model to generate a temporally grounded narrative, guiding the final generation of a temporal graph. Extensive experiments showcase the efficacy of NoT in improving various metrics. Notably, NoT attains the highest F1 on the Schema-11 evaluation set, while securing an overall F1 on par with GPT-3.5. NoT also achieves the best structural similarity across the board, even compared with GPT-3.5/4. Our code is available at https://github.com/launchnlp/NoT.
Authors: Yu Shang, Yu Li, Keyu Zhao, Likai Ma, Jiahe Liu, Fengli Xu, Yong Li
Abstract: Recent advancements in Large Language Models (LLMs) have led to a rapid growth of agentic systems capable of handling a wide range of complex tasks. However, current research largely relies on manual, task-specific design, limiting their adaptability to novel tasks. In this paper, we introduce a new research problem: Modularized LLM Agent Search (MoLAS). We propose a modular design space that abstracts existing LLM agent designs into four fundamental modules with uniform IO interface: Planning, Reasoning, Tool Use, and Memory. Building on this design space, we present a novel LLM agent search framework called AgentSquare, which introduces two core mechanisms, i.e., module evolution and recombination, to efficiently search for optimized LLM agents. To further accelerate the process, we design a performance predictor that uses in-context surrogate models to skip unpromising agent designs. Extensive experiments across six benchmarks, covering the diverse scenarios of web, embodied, tool use and game applications, show that AgentSquare substantially outperforms hand-crafted agents, achieving an average performance gain of 17.2% against best-known human designs. Moreover, AgentSquare can generate interpretable design insights, enabling a deeper understanding of agentic architecture and its impact on task performance. We believe that the modular design space and AgentSquare search framework offer a platform for fully exploiting the potential of prior successful designs and consolidating the collective efforts of research community. Code repo is available at https://github.com/tsinghua-fib-lab/AgentSquare.
Authors: Ruosen Li, Ziming Luo, Xinya Du
Abstract: Hallucinations in large language models (LLMs) pose significant challenges in tasks requiring complex multi-step reasoning, such as mathematical problem-solving. Existing approaches primarily detect the presence of hallucinations but lack a nuanced understanding of their types and manifestations. In this paper, we first introduce a comprehensive taxonomy that categorizes the common hallucinations in mathematical reasoning task into six types: fabrication, factual inconsistency, context inconsistency, instruction inconsistency, logical inconsistency, and logical error. We then propose FG-PRM (Fine-Grained Process Reward Model), an augmented model designed to detect and mitigate hallucinations in a fine-grained, step-level manner. To address the limitations of manually labeling training data, we propose an automated method for generating fine-grained hallucination data using LLMs. By injecting hallucinations into reasoning steps of correct solutions, we create a diverse and balanced synthetic dataset for training FG-PRM, which consists of six specialized Process Reward Models (PRMs), each tailored to detect a specific hallucination type. Our FG-PRM demonstrates superior performance across two key tasks: 1) Fine-grained hallucination detection: classifying hallucination types for each reasoning step; and 2) Verification: ranking multiple LLM-generated outputs to select the most accurate solution, mitigating reasoning hallucinations. Our experiments show that FG-PRM outperforms ChatGPT-3.5 and Claude-3 on fine-grained hallucination detection and substantially boosts the performance of LLMs on GSM8K and MATH benchmarks.
Authors: Runchuan Zhu, Zhipeng Ma, Jiang Wu, Junyuan Gao, Jiaqi Wang, Dahua Lin, Conghui He
Abstract: Refusal-Aware Instruction Tuning (RAIT) enables Large Language Models (LLMs) to refuse to answer unknown questions. By modifying responses of unknown questions in the training data to refusal responses such as "I don't know", RAIT enhances the reliability of LLMs and reduces their hallucination. Generally, RAIT modifies training samples based on the correctness of the initial LLM's response. However, this crude approach can cause LLMs to excessively refuse answering questions they could have correctly answered, the problem we call over-refusal. In this paper, we explore two primary causes of over-refusal: Static conflict occurs when similar samples within the LLM's feature space receive differing supervision signals (original vs. modified "I don't know"). Dynamic conflict, on the other hand, emerges as the LLM's knowledge evolves during SFT, allowing it to answer questions that were previously unanswerable. Yet, these now-answerable training samples still retain the original "I don't know" supervision signals based on the initial LLM state, resulting in inconsistencies. These conflicts cause the trained LLM to misclassify known questions as unknown, resulting in over-refusal. To address this issue, we introduce Certainty Represented Knowledge Flow for Refusal-Aware Instructions Tuning (CRaFT). CRaFT centers on two main contributions: First, we additionally incorporate response certainty to selectively filter and modify data, reducing static conflicts. Second, we implement preliminary rehearsal training to characterize changes in the LLM's knowledge state, which helps mitigate dynamic conflicts during the fine-tuning process. We conducted extensive experiments on open-ended question answering and multiple-choice question task. Experiment results show that CRaFT can improve LLM's overall performance during the RAIT process. Source code and training data will be released at Github.
Authors: Jiaqing Yuan, Ruijie Xi, Munindar P. Singh
Abstract: Argumentative stance classification plays a key role in identifying authors' viewpoints on specific topics. However, generating diverse pairs of argumentative sentences across various domains is challenging. Existing benchmarks often come from a single domain or focus on a limited set of topics. Additionally, manual annotation for accurate labeling is time-consuming and labor-intensive. To address these challenges, we propose leveraging platform rules, readily available expert-curated content, and large language models to bypass the need for human annotation. Our approach produces a multidomain benchmark comprising 4,498 topical claims and 30,961 arguments from three sources, spanning 21 domains. We benchmark the dataset in fully supervised, zero-shot, and few-shot settings, shedding light on the strengths and limitations of different methodologies. We release the dataset and code in this study at hidden for anonymity.
Authors: Siyi Liu, Qiang Ning, Kishaloy Halder, Wei Xiao, Zheng Qi, Phu Mon Htut, Yi Zhang, Neha Anna John, Bonan Min, Yassine Benajiba, Dan Roth
Abstract: Open domain question answering systems frequently rely on information retrieved from large collections of text (such as the Web) to answer questions. However, such collections of text often contain conflicting information, and indiscriminately depending on this information may result in untruthful and inaccurate answers. To understand the gravity of this problem, we collect a human-annotated dataset, Question Answering with Conflicting Contexts (QACC), and find that as much as 25% of unambiguous, open domain questions can lead to conflicting contexts when retrieved using Google Search. We evaluate and benchmark three powerful Large Language Models (LLMs) with our dataset QACC and demonstrate their limitations in effectively addressing questions with conflicting information. To explore how humans reason through conflicting contexts, we request our annotators to provide explanations for their selections of correct answers. We demonstrate that by finetuning LLMs to explain their answers, we can introduce richer information into their training that guide them through the process of reasoning with conflicting contexts.
Authors: Xiongtao Zhou, Jie He, Lanyu Chen, Jingyu Li, Haojing Chen, V\'ictor Guti\'errez-Basulto, Jeff Z. Pan, Hanjie Chen
Abstract: Multimodal Chain of Thought (MCoT) is a popular prompting strategy for improving the performance of multimodal large language models (MLLMs) across a range of complex reasoning tasks. Despite its popularity, there is a notable absence of automated methods for evaluating the quality of reasoning steps in MCoT. To address this gap, we propose Multimodal Chain-of-Thought Evaluation (MiCEval), a framework designed to assess the correctness of reasoning chains by evaluating the quality of both the description and each reasoning step. The evaluation of the description component focuses on the accuracy of the image descriptions, while the reasoning step evaluates the quality of each step as it is conditionally generated based on the preceding steps. MiCEval is built upon a fine-grained dataset with annotations that rate each step according to correctness, relevance, and informativeness. Extensive experiments on four state-of-the-art MLLMs show that step-wise evaluations using MiCEval align more closely with human judgments compared to existing methods based on cosine similarity or fine-tuning approaches. MiCEval datasets and code can be found in https://github.com/alenai97/MiCEval.
Authors: Jing Yang Lee, Seokhwan Kim, Kartik Mehta, Jiun-Yu Kao, Yu-Hsiang Lin, Arpit Gupta
Abstract: Information-Seeking Dialogue (ISD) agents aim to provide accurate responses to user queries. While proficient in directly addressing user queries, these agents, as well as LLMs in general, predominantly exhibit reactive behavior, lacking the ability to generate proactive responses that actively engage users in sustained conversations. However, existing definitions of proactive dialogue in this context do not focus on how each response actively engages the user and sustains the conversation. Hence, we present a new definition of proactivity that focuses on enhancing the `proactiveness' of each generated response via the introduction of new information related to the initial query. To this end, we construct a proactive dialogue dataset comprising 2,000 single-turn conversations, and introduce several automatic metrics to evaluate response `proactiveness' which achieved high correlation with human annotation. Additionally, we introduce two innovative Chain-of-Thought (CoT) prompts, the 3-step CoT and the 3-in-1 CoT prompts, which consistently outperform standard prompts by up to 90% in the zero-shot setting.
Authors: Catherine Arnett, Eliot Jones, Ivan P. Yamshchikov, Pierre-Carl Langlais
Abstract: Open-source large language models are becoming increasingly available and popular among researchers and practitioners. While significant progress has been made on open-weight models, open training data is a practice yet to be adopted by the leading open-weight models creators. At the same time, there researchers are working to make language models safer. We propose a data curation pipeline to reduce harmful outputs by models trained on public domain data. There are unique challenges to working with public domain data, as these sources differ from web text in both form and content. Many sources are historical documents and are the result of Optical Character Recognition (OCR). Consequently, current state-of-the-art approaches to toxicity filtering are often infeasible or inappropriate for open data models. In this paper, we introduce a new fully open-source pipeline for open-data toxicity filtering. Our contributions are threefold. We create a custom training dataset, ToxicCommons, which is composed of texts which have been classified across five different dimensions (racial/origin-based, gender/sex-based, religious, ability-based discrimination, and violence). We use this dataset to train a custom classifier, Celadon, that can be used to detect toxic content in open data more efficiently at a larger scale. Finally, we describe the balanced approach to content filtration that optimizes safety filtering with respect to the filtered data available for training.
Authors: Zifeng Wang, Hanyin Wang, Benjamin Danek, Ying Li, Christina Mack, Hoifung Poon, Yajuan Wang, Pranav Rajpurkar, Jimeng Sun
Abstract: The integration of Large Language Models (LLMs) into medical applications has sparked widespread interest across the healthcare industry, from drug discovery and development to clinical decision support, assisting telemedicine, medical devices, and healthcare insurance applications. This perspective paper aims to discuss the inner workings of building LLM-powered medical AI applications and introduces a comprehensive framework for their development. We review existing literature and outline the unique challenges of applying LLMs in specialized medical contexts. Additionally, we introduce a three-step framework to organize medical LLM research activities: 1) Modeling: breaking down complex medical workflows into manageable steps for developing medical-specific models; 2) Optimization: optimizing the model performance with crafted prompts and integrating external knowledge and tools, and 3) System engineering: decomposing complex tasks into subtasks and leveraging human expertise for building medical AI applications. Furthermore, we offer a detailed use case playbook that describes various LLM-powered medical AI applications, such as optimizing clinical trial design, enhancing clinical decision support, and advancing medical imaging analysis. Finally, we discuss various challenges and considerations for building medical AI applications with LLMs, such as handling hallucination issues, data ownership and compliance, privacy, intellectual property considerations, compute cost, sustainability issues, and responsible AI requirements.
Authors: Amogh Akella
Abstract: In this paper, we investigate how to harness large language models (LLMs) to solve mathematical problems both quickly and accurately. Specifically, we demonstrate the effectiveness of classifying problems into distinct categories and applying category-specific problem-solving strategies to enhance the math performance of LLMs. We develop a straightforward machine learning model for problem categorization and show that its accuracy can be significantly improved through the creation of well-designed training datasets. We believe that our approach works by helping reduce hallucinations in LLMs, which is a critical step toward unlocking their potential to tackle advanced mathematical problems.
Authors: Gideon Yoffe, Nachum Dershowitz, Ariel Vishne, Barak Sober
Abstract: Stylometry aims to distinguish authors by analyzing literary traits assumed to reflect semi-conscious choices distinct from elements like genre or theme. However, these components often overlap, complicating text classification based solely on feature distributions. While some literary properties, such as thematic content, are likely to manifest as correlations between adjacent text units, others, like authorial style, may be independent thereof. We introduce a hypothesis-testing approach to evaluate the influence of sequentially correlated literary properties on text classification, aiming to determine when these correlations drive classification. Using a multivariate binary distribution, our method models sequential correlations between text units as a stochastic process, assessing the likelihood of clustering across varying adjacency scales. This enables us to examine whether classification is dominated by sequentially correlated properties or remains independent. In experiments on a diverse English prose corpus, our analysis integrates traditional and neural embeddings within supervised and unsupervised frameworks. Results demonstrate that our approach effectively identifies when textual classification is not primarily influenced by sequentially correlated literary properties, particularly in cases where texts differ in authorial style or genre rather than by a single author within a similar genre.
Authors: Israel Abebe Azime, Atnafu Lambebo Tonja, Tadesse Destaw Belay, Yonas Chanie, Bontu Fufa Balcha, Negasi Haile Abadi, Henok Biadglign Ademtew, Mulubrhan Abebe Nerea, Debela Desalegn Yadeta, Derartu Dagne Geremew, Assefa Atsbiha tesfau, Philipp Slusallek, Thamar Solorio, Dietrich Klakow
Abstract: With the rapid development of evaluation datasets to assess LLMs understanding across a wide range of subjects and domains, identifying a suitable language understanding benchmark has become increasingly challenging. In this work, we explore LLM evaluation challenges for low-resource language understanding and introduce ProverbEval, LLM evaluation benchmark for low-resource languages based on proverbs to focus on low-resource language understanding in culture-specific scenarios. We benchmark various LLMs and explore factors that create variability in the benchmarking process. We observed performance variances of up to 50%, depending on the order in which answer choices were presented in multiple-choice tasks. Native language proverb descriptions significantly improve tasks such as proverb generation, contributing to improved outcomes. Additionally, monolingual evaluations consistently outperformed their cross-lingual counterparts. We argue special attention must be given to the order of choices, choice of prompt language, task variability, and generation tasks when creating LLM evaluation benchmarks.
Authors: Zijian Hu, Jipeng Zhang, Rui Pan, Zhaozhuo Xu, Shanshan Han, Han Jin, Alay Dilipbhai Shah, Dimitris Stripelis, Yuhang Yao, Salman Avestimehr, Chaoyang He, Tong Zhang
Abstract: We present Fox-1, a series of small language models (SLMs) consisting of Fox-1-1.6B and Fox-1-1.6B-Instruct-v0.1. These models are pre-trained on 3 trillion tokens of web-scraped document data and fine-tuned with 5 billion tokens of instruction-following and multi-turn conversation data. Aiming to improve the pre-training efficiency, Fox-1-1.6B model introduces a novel 3-stage data curriculum across all the training data with 2K-8K sequence length. In architecture design, Fox-1 features a deeper layer structure, an expanded vocabulary, and utilizes Grouped Query Attention (GQA), offering a performant and efficient architecture compared to other SLMs. Fox-1 achieves better or on-par performance in various benchmarks compared to StableLM-2-1.6B, Gemma-2B, Qwen1.5-1.8B, and OpenELM1.1B, with competitive inference speed and throughput. The model weights have been released under the Apache 2.0 license, where we aim to promote the democratization of LLMs and make them fully accessible to the whole open-source community.
Authors: Anton Alekseev, Timur Turatali
Abstract: Large language models (LLMs) have excelled in numerous benchmarks, advancing AI applications in both linguistic and non-linguistic tasks. However, this has primarily benefited well-resourced languages, leaving less-resourced ones (LRLs) at a disadvantage. In this paper, we highlight the current state of the NLP field in the specific LRL: kyrgyz tili. Human evaluation, including annotated datasets created by native speakers, remains an irreplaceable component of reliable NLP performance, especially for LRLs where automatic evaluations can fall short. In recent assessments of the resources for Turkic languages, Kyrgyz is labeled with the status 'Scraping By', a severely under-resourced language spoken by millions. This is concerning given the growing importance of the language, not only in Kyrgyzstan but also among diaspora communities where it holds no official status. We review prior efforts in the field, noting that many of the publicly available resources have only recently been developed, with few exceptions beyond dictionaries (the processed data used for the analysis is presented at https://kyrgyznlp.github.io/). While recent papers have made some headway, much more remains to be done. Despite interest and support from both business and government sectors in the Kyrgyz Republic, the situation for Kyrgyz language resources remains challenging. We stress the importance of community-driven efforts to build these resources, ensuring the future advancement sustainability. We then share our view of the most pressing challenges in Kyrgyz NLP. Finally, we propose a roadmap for future development in terms of research topics and language resources.
Authors: Zhidong Ling, Zihao Li, Pablo Romero, Lifeng Han, Goran Nenadic
Abstract: This report is the system description of the BeeManc team for shared task Plain Language Adaptation of Biomedical Abstracts (PLABA) 2024. This report contains two sections corresponding to the two sub-tasks in PLABA 2024. In task one, we applied fine-tuned ReBERTa-Base models to identify and classify the difficult terms, jargon and acronyms in the biomedical abstracts and reported the F1 score. Due to time constraints, we didn't finish the replacement task. In task two, we leveraged Llamma3.1-70B-Instruct and GPT-4o with the one-shot prompts to complete the abstract adaptation and reported the scores in BLEU, SARI, BERTScore, LENS, and SALSA. From the official Evaluation from PLABA-2024 on Task 1A and 1B, our \textbf{much smaller fine-tuned RoBERTa-Base} model ranked 3rd and 2nd respectively on the two sub-task, and the \textbf{1st on averaged F1 scores across the two tasks} from 9 evaluated systems. Our LLaMA-3.1-70B-instructed model achieved the \textbf{highest Completeness} score for Task-2. We share our fine-tuned models and related resources at \url{https://github.com/HECTA-UoM/PLABA2024}
Authors: Cl\'ement Dumas, Chris Wendler, Veniamin Veselovsky, Giovanni Monea, Robert West
Abstract: A central question in multilingual language modeling is whether large language models (LLMs) develop a universal concept representation, disentangled from specific languages. In this paper, we address this question by analyzing latent representations (latents) during a word translation task in transformer-based LLMs. We strategically extract latents from a source translation prompt and insert them into the forward pass on a target translation prompt. By doing so, we find that the output language is encoded in the latent at an earlier layer than the concept to be translated. Building on this insight, we conduct two key experiments. First, we demonstrate that we can change the concept without changing the language and vice versa through activation patching alone. Second, we show that patching with the mean over latents across different languages does not impair and instead improves the models' performance in translating the concept. Our results provide evidence for the existence of language-agnostic concept representations within the investigated models.
Authors: Yan Hu, Xu Zuo, Yujia Zhou, Xueqing Peng, Jimin Huang, Vipina K. Keloth, Vincent J. Zhang, Ruey-Ling Weng, Qingyu Chen, Xiaoqian Jiang, Kirk E. Roberts, Hua Xu
Abstract: Backgrounds: Information extraction (IE) is critical in clinical natural language processing (NLP). While large language models (LLMs) excel on generative tasks, their performance on extractive tasks remains debated. Methods: We investigated Named Entity Recognition (NER) and Relation Extraction (RE) using 1,588 clinical notes from four sources (UT Physicians, MTSamples, MIMIC-III, and i2b2). We developed an annotated corpus covering 4 clinical entities and 16 modifiers, and compared instruction-tuned LLaMA-2 and LLaMA-3 against BiomedBERT in terms of performance, generalizability, computational resources, and throughput to BiomedBERT. Results: LLaMA models outperformed BiomedBERT across datasets. With sufficient training data, LLaMA showed modest improvements (1% on NER, 1.5-3.7% on RE); improvements were larger with limited training data. On unseen i2b2 data, LLaMA-3-70B outperformed BiomedBERT by 7% (F1) on NER and 4% on RE. However, LLaMA models required more computing resources and ran up to 28 times slower. We implemented "Kiwi," a clinical IE package featuring both models, available at https://kiwi.clinicalnlp.org/. Conclusion: This study is among the first to develop and evaluate a comprehensive clinical IE system using open-source LLMs. Results indicate that LLaMA models outperform BiomedBERT for clinical NER and RE but with higher computational costs and lower throughputs. These findings highlight that choosing between LLMs and traditional deep learning methods for clinical IE applications should remain task-specific, taking into account both performance metrics and practical considerations such as available computing resources and the intended use case scenarios.
Authors: Deqing Fu, Tian-Qi Chen, Robin Jia, Vatsal Sharan
Abstract: Transformers excel at in-context learning (ICL) -- learning from demonstrations without parameter updates -- but how they do so remains a mystery. Recent work suggests that Transformers may internally run Gradient Descent (GD), a first-order optimization method, to perform ICL. In this paper, we instead demonstrate that Transformers learn to approximate second-order optimization methods for ICL. For in-context linear regression, Transformers share a similar convergence rate as Iterative Newton's Method, both exponentially faster than GD. Empirically, predictions from successive Transformer layers closely match different iterations of Newton's Method linearly, with each middle layer roughly computing 3 iterations; thus, Transformers and Newton's method converge at roughly the same rate. In contrast, Gradient Descent converges exponentially more slowly. We also show that Transformers can learn in-context on ill-conditioned data, a setting where Gradient Descent struggles but Iterative Newton succeeds. Finally, to corroborate our empirical findings, we prove that Transformers can implement $k$ iterations of Newton's method with $k + \mathcal{O}(1)$ layers.
Authors: Trilok Padhi, Ugur Kursuncu, Yaman Kumar, Valerie L. Shalin, Lane Peterson Fronczek
Abstract: The digital landscape continually evolves with multimodality, enriching the online experience for users. Creators and marketers aim to weave subtle contextual cues from various modalities into congruent content to engage users with a harmonious message. This interplay of multimodal cues is often a crucial factor in attracting users' attention. However, this richness of multimodality presents a challenge to computational modeling, as the semantic contextual cues spanning across modalities need to be unified to capture the true holistic meaning of the multimodal content. This contextual meaning is critical in attracting user engagement as it conveys the intended message of the brand or the organization. In this work, we incorporate external commonsense knowledge from knowledge graphs to enhance the representation of multimodal data using compact Visual Language Models (VLMs) and predict the success of multi-modal crowdfunding campaigns. Our results show that external knowledge commonsense bridges the semantic gap between text and image modalities, and the enhanced knowledge-infused representations improve the predictive performance of models for campaign success upon the baselines without knowledge. Our findings highlight the significance of contextual congruence in online multimodal content for engaging and successful crowdfunding campaigns.
Authors: Zhengyuan Jiang, Moyang Guo, Yuepeng Hu, Neil Zhenqiang Gong
Abstract: Several companies have deployed watermark-based detection to identify AI-generated content. However, attribution--the ability to trace back to the user of a generative AI (GenAI) service who created a given piece of AI-generated content--remains largely unexplored despite its growing importance. In this work, we aim to bridge this gap by conducting the first systematic study on watermark-based, user-level attribution of AI-generated content. Our key idea is to assign a unique watermark to each user of the GenAI service and embed this watermark into the AI-generated content created by that user. Attribution is then performed by identifying the user whose watermark best matches the one extracted from the given content. This approach, however, faces a key challenge: How should watermarks be selected for users to maximize attribution performance? To address the challenge, we first theoretically derive lower bounds on detection and attribution performance through rigorous probabilistic analysis for any given set of user watermarks. Then, we select watermarks for users to maximize these lower bounds, thereby optimizing detection and attribution performance. Our theoretical and empirical results show that watermark-based attribution inherits both the accuracy and (non-)robustness properties of the underlying watermark. Specifically, attribution remains highly accurate when the watermarked AI-generated content is either not post-processed or subjected to common post-processing such as JPEG compression, as well as black-box adversarial post-processing with limited query budgets.
Authors: Yifei Wang, Yuyang Wu, Zeming Wei, Stefanie Jegelka, Yisen Wang
Abstract: Going beyond mimicking limited human experiences, recent studies show initial evidence that, like humans, large language models (LLMs) are capable of improving their abilities purely by self-correction, i.e., correcting previous responses through self-examination, in certain circumstances. Nevertheless, little is known about how such capabilities arise. In this work, based on a simplified setup akin to an alignment task, we theoretically analyze self-correction from an in-context learning perspective, showing that when LLMs give relatively accurate self-examinations as rewards, they are capable of refining responses in an in-context way. Notably, going beyond previous theories on over-simplified linear transformers, our theoretical construction underpins the roles of several key designs of realistic transformers for self-correction: softmax attention, multi-head attention, and the MLP block. We validate these findings extensively on synthetic datasets. Inspired by these findings, we also illustrate novel applications of self-correction, such as defending against LLM jailbreaks, where a simple self-correction step does make a large difference. We believe that these findings will inspire further research on understanding, exploiting, and enhancing self-correction for building better foundation models.
Authors: JiaYing Zheng, HaiNan Zhang, LingXiang Wang, WangJie Qiu, HongWei Zheng, ZhiMing Zheng
Abstract: Private data, being larger and quality-higher than public data, can greatly improve large language models (LLM). However, due to privacy concerns, this data is often dispersed in multiple silos, making its secure utilization for LLM training a challenge. Federated learning (FL) is an ideal solution for training models with distributed private data, but traditional frameworks like FedAvg are unsuitable for LLM due to their high computational demands on clients. An alternative, split learning, offloads most training parameters to the server while training embedding and output layers locally, making it more suitable for LLM. Nonetheless, it faces significant challenges in security and efficiency. Firstly, the gradients of embeddings are prone to attacks, leading to potential reverse engineering of private data. Furthermore, the server's limitation of handle only one client's training request at a time hinders parallel training, severely impacting training efficiency. In this paper, we propose a Federated Learning framework for LLM, named FL-GLM, which prevents data leakage caused by both server-side and peer-client attacks while improving training efficiency. Specifically, we first place the input block and output block on local client to prevent embedding gradient attacks from server. Secondly, we employ key-encryption during client-server communication to prevent reverse engineering attacks from peer-clients. Lastly, we employ optimization methods like client-batching or server-hierarchical, adopting different acceleration methods based on the actual computational capabilities of the server. Experimental results on NLU and generation tasks demonstrate that FL-GLM achieves comparable metrics to centralized chatGLM model, validating the effectiveness of our federated learning framework.
Authors: Sukmin Yun, Haokun Lin, Rusiru Thushara, Mohammad Qazim Bhat, Yongxin Wang, Zutao Jiang, Mingkai Deng, Jinhong Wang, Tianhua Tao, Junbo Li, Haonan Li, Preslav Nakov, Timothy Baldwin, Zhengzhong Liu, Eric P. Xing, Xiaodan Liang, Zhiqiang Shen
Abstract: Multimodal large language models (MLLMs) have shown impressive success across modalities such as image, video, and audio in a variety of understanding and generation tasks. However, current MLLMs are surprisingly poor at understanding webpage screenshots and generating their corresponding HTML code. To address this problem, we propose $\texttt{Web2Code}$, a benchmark consisting of a new large-scale webpage-to-code dataset for instruction tuning and an evaluation framework for the webpage understanding and HTML code translation abilities of MLLMs. For dataset construction, we leverage pretrained LLMs to enhance existing webpage-to-code datasets as well as generate a diverse pool of new webpages rendered into images. Specifically, the inputs are webpage images and instructions, while the responses are the webpage's HTML code. We further include diverse natural language QA pairs about the webpage content in the responses to enable a more comprehensive understanding of the web content. To evaluate model performance in these tasks, we develop an evaluation framework for testing MLLMs' abilities in webpage understanding and web-to-code generation. Extensive experiments show that our proposed dataset is beneficial not only to our proposed tasks but also in the general visual domain. We hope our work will contribute to the development of general MLLMs suitable for web-based content generation and task automation. Our data and code are available at https://github.com/MBZUAI-LLM/web2code.
Authors: Philipp Allgeuer, Kyra Ahrens, Stefan Wermter
Abstract: We introduce NOVIC, an innovative real-time uNconstrained Open Vocabulary Image Classifier that uses an autoregressive transformer to generatively output classification labels as language. Leveraging the extensive knowledge of CLIP models, NOVIC harnesses the embedding space to enable zero-shot transfer from pure text to images. Traditional CLIP models, despite their ability for open vocabulary classification, require an exhaustive prompt of potential class labels, restricting their application to images of known content or context. To address this, we propose an "object decoder" model that is trained on a large-scale 92M-target dataset of templated object noun sets and LLM-generated captions to always output the object noun in question. This effectively inverts the CLIP text encoder and allows textual object labels from essentially the entire English language to be generated directly from image-derived embedding vectors, without requiring any a priori knowledge of the potential content of an image, and without any label biases. The trained decoders are tested on a mix of manually and web-curated datasets, as well as standard image classification benchmarks, and achieve fine-grained prompt-free prediction scores of up to 87.5%, a strong result considering the model must work for any conceivable image and without any contextual clues.
Authors: Liana Patel, Siddharth Jha, Parth Asawa, Melissa Pan, Carlos Guestrin, Matei Zaharia
Abstract: The semantic capabilities of language models (LMs) have the potential to enable rich analytics and reasoning over vast knowledge corpora. Unfortunately, existing systems lack high-level abstractions to perform bulk semantic queries across large corpora. We introduce semantic operators, a declarative programming interface that extends the relational model with composable AI-based operations for bulk semantic queries (e.g., filtering, sorting, joining or aggregating records using natural language criteria). Each operator can be implemented and optimized in multiple ways, opening a rich space for execution plans similar to relational operators. We implement our operators in LOTUS, an open source query engine with a DataFrame API. Furthermore, we develop several novel optimizations that take advantage of the declarative nature of semantic operators to accelerate semantic filtering, clustering and join operators by up to $400\times$ while offering statistical accuracy guarantees. We demonstrate LOTUS' effectiveness on real AI applications including fact-checking, extreme multi-label classification, and search. We show that the semantic operator model is expressive, capturing state-of-the-art AI pipelines in a few operator calls, and making it easy to express new pipelines that achieve up to $180\%$ higher quality. Overall, LOTUS queries match or exceed the accuracy of state-of-the-art AI pipelines for each task while running up to 28$\times$ faster. LOTUS is publicly available at https://github.com/stanford-futuredata/lotus.
Authors: Md Zarif Hossain, Ahmed Imteaj
Abstract: Vision-language models (VLMs) have achieved significant strides in recent times specially in multimodal tasks, yet they remain susceptible to adversarial attacks on their vision components. To address this, we propose Sim-CLIP, an unsupervised adversarial fine-tuning method that enhances the robustness of the widely-used CLIP vision encoder against such attacks while maintaining semantic richness and specificity. By employing a Siamese architecture with cosine similarity loss, Sim-CLIP learns semantically meaningful and attack-resilient visual representations without requiring large batch sizes or momentum encoders. Our results demonstrate that VLMs enhanced with Sim-CLIP's fine-tuned CLIP encoder exhibit significantly enhanced robustness against adversarial attacks, while preserving semantic meaning of the perturbed images. Notably, Sim-CLIP does not require additional training or fine-tuning of the VLM itself; replacing the original vision encoder with our fine-tuned Sim-CLIP suffices to provide robustness. This work underscores the significance of reinforcing foundational models like CLIP to safeguard the reliability of downstream VLM applications, paving the way for more secure and effective multimodal systems.
Authors: Fuheng Zhao, Divyakant Agrawal, Amr El Abbadi
Abstract: Database queries traditionally operate under the closed-world assumption, providing no answers to questions that require information beyond the data stored in the database. Hybrid querying using SQL offers an alternative by integrating relational databases with large language models (LLMs) to answer beyond-database questions. In this paper, we present the first cross-domain benchmark, SWAN, containing 120 beyond-database questions over four real-world databases. To leverage state-of-the-art language models in addressing these complex questions in SWAN, we present two solutions: one based on schema expansion and the other based on user defined functions. We also discuss optimization opportunities and potential future directions. Our evaluation demonstrates that using GPT-4 Turbo with few-shot prompts, one can achieves up to 40.0\% in execution accuracy and 48.2\% in data factuality. These results highlights both the potential and challenges for hybrid querying. We believe that our work will inspire further research in creating more efficient and accurate data systems that seamlessly integrate relational databases and large language models to address beyond-database questions.
Authors: Qiang Sun, Yuanyi Luo, Sirui Li, Wenxiao Zhang, Wei Liu
Abstract: Multimodal conversational agents are highly desirable because they offer natural and human-like interaction. However, there is a lack of comprehensive end-to-end solutions to support collaborative development and benchmarking. While proprietary systems like GPT-4o and Gemini demonstrating impressive integration of audio, video, and text with response times of 200-250ms, challenges remain in balancing latency, accuracy, cost, and data privacy. To better understand and quantify these issues, we developed OpenOmni, an open-source, end-to-end pipeline benchmarking tool that integrates advanced technologies such as Speech-to-Text, Emotion Detection, Retrieval Augmented Generation, Large Language Models, along with the ability to integrate customized models. OpenOmni supports local and cloud deployment, ensuring data privacy and supporting latency and accuracy benchmarking. This flexible framework allows researchers to customize the pipeline, focusing on real bottlenecks and facilitating rapid proof-of-concept development. OpenOmni can significantly enhance applications like indoor assistance for visually impaired individuals, advancing human-computer interaction. Our demonstration video is available https://www.youtube.com/watch?v=zaSiT3clWqY, demo is available via https://openomni.ai4wa.com, code is available via https://github.com/AI4WA/OpenOmniFramework.
URLs: https://www.youtube.com/watch?v=zaSiT3clWqY,, https://openomni.ai4wa.com,, https://github.com/AI4WA/OpenOmniFramework.
Authors: Qian Niu, Junyu Liu, Ziqian Bi, Pohsun Feng, Benji Peng, Keyu Chen, Ming Li, Lawrence KQ Yan, Yichao Zhang, Caitlyn Heqi Yin, Cheng Fei, Tianyang Wang, Yunze Wang, Silin Chen
Abstract: This comprehensive review explores the intersection of Large Language Models (LLMs) and cognitive science, examining similarities and differences between LLMs and human cognitive processes. We analyze methods for evaluating LLMs cognitive abilities and discuss their potential as cognitive models. The review covers applications of LLMs in various cognitive fields, highlighting insights gained for cognitive science research. We assess cognitive biases and limitations of LLMs, along with proposed methods for improving their performance. The integration of LLMs with cognitive architectures is examined, revealing promising avenues for enhancing artificial intelligence (AI) capabilities. Key challenges and future research directions are identified, emphasizing the need for continued refinement of LLMs to better align with human cognition. This review provides a balanced perspective on the current state and future potential of LLMs in advancing our understanding of both artificial and human intelligence.
Authors: Andrei Chernov
Abstract: Inspired by recent advancements in large language models (LLMs) for Natural Language Processing (NLP), there has been a surge in research focused on developing foundational models for time series forecasting. One approach involves training LLM architectures on tokenized time series data using cross-entropy loss. Although this method has demonstrated promising results, cross-entropy loss is primarily designed for classification tasks and does not account for the distance between classes. To address this limitation, we propose using the Wasserstein loss for such architectures. To validate our approach, we fine-tuned a foundational time series model on $22$ zero-shot datasets, comparing the performance of cross-entropy loss with that of Wasserstein loss. Our results demonstrate that replacing cross-entropy loss with Wasserstein loss significantly improves point estimation.
Authors: Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz, Yiming Xiao
Abstract: Segmentation of anatomical structures and pathological regions in medical images is essential for modern clinical diagnosis, disease research, and treatment planning. While significant advancements have been made in deep learning-based segmentation techniques, many of these methods still suffer from limitations in data efficiency, generalizability, and interactivity. As a result, developing precise segmentation methods that require fewer labeled datasets remains a critical challenge in medical image analysis. Recently, the introduction of foundation models like CLIP and Segment-Anything-Model (SAM), with robust cross-domain representations, has paved the way for interactive and universal image segmentation. However, further exploration of these models for data-efficient segmentation in medical imaging is still needed and highly relevant. In this paper, we introduce MedCLIP-SAMv2, a novel framework that integrates the CLIP and SAM models to perform segmentation on clinical scans using text prompts, in both zero-shot and weakly supervised settings. Our approach includes fine-tuning the BiomedCLIP model with a new Decoupled Hard Negative Noise Contrastive Estimation (DHN-NCE) loss, and leveraging the Multi-modal Information Bottleneck (M2IB) to create visual prompts for generating segmentation masks from SAM in the zero-shot setting. We also investigate using zero-shot segmentation labels within a weakly supervised paradigm to enhance segmentation quality further. Extensive testing across four diverse segmentation tasks and medical imaging modalities (breast tumor ultrasound, brain tumor MRI, lung X-ray, and lung CT) demonstrates the high accuracy of our proposed framework. Our code is available at https://github.com/HealthX-Lab/MedCLIP-SAMv2.
Authors: Xinfeng Wang, Jin Cui, Fumiyo Fukumoto, Yoshimi Suzuki
Abstract: Large language models (LLMs) have demonstrated prominent reasoning capabilities in recommendation tasks by transforming them into text-generation tasks. However, existing approaches either disregard or ineffectively model the user-item high-order interactions. To this end, this paper presents an enhanced LLM-based recommender (ELMRec). We enhance whole-word embeddings to substantially enhance LLMs' interpretation of graph-constructed interactions for recommendations, without requiring graph pre-training. This finding may inspire endeavors to incorporate rich knowledge graphs into LLM-based recommenders via whole-word embedding. We also found that LLMs often recommend items based on users' earlier interactions rather than recent ones, and present a reranking solution. Our ELMRec outperforms state-of-the-art (SOTA) methods in both direct and sequential recommendations.
Authors: Tonghe Zhuang, Zhicheng Lin
Abstract: Computer programming (coding) is indispensable for researchers across disciplines, yet it remains challenging to learn and time-consuming to carry out. Generative AI, particularly large language models (LLMs), has the potential to transform coding into intuitive conversations, but best practices and effective workflows are only emerging. We dissect AI-based coding through three key lenses: the nature and role of LLMs in coding (why), six types of coding assistance they provide (what), and a five-step workflow in action with practical implementation strategies (how). Additionally, we address the limitations and future outlook of AI in coding. By offering actionable insights, this framework helps to guide researchers in effectively leveraging AI to enhance coding practices and education, accelerating scientific progress.
Authors: Chenhang Cui, An Zhang, Yiyang Zhou, Zhaorun Chen, Gelei Deng, Huaxiu Yao, Tat-Seng Chua
Abstract: The recent advancements in large language models (LLMs) and pre-trained vision models have accelerated the development of vision-language large models (VLLMs), enhancing the interaction between visual and linguistic modalities. Despite their notable success across various domains, VLLMs face challenges in modality alignment, which can lead to issues like hallucinations and unsafe content generation. Current alignment techniques often rely on coarse feedback and external datasets, limiting scalability and performance. In this paper, we propose FiSAO (Fine-Grained Self-Alignment Optimization), a novel self-alignment method that utilizes the model's own visual encoder as a fine-grained verifier to improve vision-language alignment without the need for additional data. By leveraging token-level feedback from the vision encoder, FiSAO significantly improves vision-language alignment, even surpassing traditional preference tuning methods that require additional data. Through both theoretical analysis and experimental validation, we demonstrate that FiSAO effectively addresses the misalignment problem in VLLMs, marking the first instance of token-level rewards being applied to such models.
Authors: Yuqing Zhou, Ziwei Zhu
Abstract: In text classification tasks, models often rely on spurious correlations for predictions, incorrectly associating irrelevant features with the target labels. This issue limits the robustness and generalization of models, especially when faced with out-of-distribution data where such spurious correlations no longer hold. To address this challenge, we propose the Causally Calibrated Robust Classifier (CCR), which aims to reduce models' reliance on spurious correlations and improve model robustness. Our approach integrates a causal feature selection method based on counterfactual reasoning, along with an unbiased inverse propensity weighting (IPW) loss function. By focusing on selecting causal features, we ensure that the model relies less on spurious features during prediction. We theoretically justify our approach and empirically show that CCR achieves state-of-the-art performance among methods without group labels, and in some cases, it can compete with the models that utilize group labels.
Authors: Wen-Ding Li, Keya Hu, Carter Larsen, Yuqing Wu, Simon Alford, Caleb Woo, Spencer M. Dunn, Hao Tang, Michelangelo Naim, Dat Nguyen, Wei-Long Zheng, Zenna Tavares, Yewen Pu, Kevin Ellis
Abstract: When learning an input-output mapping from very few examples, is it better to first infer a latent function that explains the examples, or is it better to directly predict new test outputs, e.g. using a neural network? We study this question on ARC, a highly diverse dataset of abstract reasoning tasks. We train neural models for induction (inferring latent functions) and transduction (directly predicting the test output for a given test input). Our models are trained on synthetic data generated by prompting LLMs to produce Python code specifying a function to be inferred, plus a stochastic subroutine for generating inputs to that function. We find inductive and transductive models solve very different problems, despite training on the same problems, and despite sharing the same neural architecture.
Authors: Siraj Munir, Alessandro Aldini
Abstract: Large Language Models (LLMs) are revolutionizing the landscape of Generative Artificial Intelligence (GenAI), with innovative LLM-backed solutions emerging rapidly. However, when applied to database technologies, specifically query generation for graph databases and Knowledge Graphs (KGs), LLMs still face significant challenges. While research on LLM-driven query generation for Structured Query Language (SQL) exists, similar systems for graph databases remain underdeveloped. This paper presents a comparative study addressing the challenge of generating Cypher queries a powerful language for interacting with graph databases using open-access LLMs. We rigorously evaluate several LLM agents (OpenAI ChatGPT 4o, Claude Sonnet 3.5, Google Gemini Pro 1.5, and a locally deployed Llama 3.1 8B) using a designed few-shot learning prompt and Retrieval Augmented Generation (RAG) backed by Chain-of-Thoughts (CoT) reasoning. Our empirical analysis of query generation accuracy reveals that Claude Sonnet 3.5 outperforms its counterparts in this specific domain. Further, we highlight promising future research directions to address the identified limitations and advance LLM-driven query generation for graph databases.
Authors: Abhinav Java, Simra Shahid, Chirag Agarwal
Abstract: The widespread practice of indiscriminate data scraping to fine-tune language models (LMs) raises significant legal and ethical concerns, particularly regarding compliance with data protection laws such as the General Data Protection Regulation (GDPR). This practice often results in the unauthorized use of personal information, prompting growing debate within the academic and regulatory communities. Recent works have introduced the concept of generating unlearnable datasets (by adding imperceptible noise to the clean data), such that the underlying model achieves lower loss during training but fails to generalize to the unseen test setting. Though somewhat effective, these approaches are predominantly designed for images and are limited by several practical constraints like requiring knowledge of the target model. To this end, we introduce RegText, a framework that injects imperceptible spurious correlations into natural language datasets, effectively rendering them unlearnable without affecting semantic content. We demonstrate RegText's utility through rigorous empirical analysis of small and large LMs. Notably, RegText can restrict newer models like GPT-4o and Llama from learning on our generated data, resulting in a drop in their test accuracy compared to their zero-shot performance and paving the way for generating unlearnable text to protect public data.