论文检索

Zefei Long, Zhenbiao Cao, Wei Chen, Zhongyu Wei

Electromyography (EMG) tables are crucial for diagnosing muscle and nerve disorders, and advancing the automation of EMG diagnostics is significant for improving medical efficiency. EMG tables contain extensive continuous numerical data, which current Large Language Models (LLMs) often struggle to interpret effectively. To address this issue, we propose EMGLLM, a data-to-text model specifically designed for medical examination tables. EMGLLM employs the EMG Alignment Encoder to simulate the process that doctors compare test values with reference values, aligning the data into word embeddings that reflect health degree. Additionally, we construct ETM, a dataset comprising 17,250 real cases and their corresponding diagnostic results, to support medical data-to-text tasks. Experimental results on ETM demonstrate that EMGLLM outperforms various baseline models in understanding EMG tables and generating high-quality diagnoses, which represents an effective paradigm for automatic diagnosis generation from medical examination table.

Xinzhe Zheng, Sijie Ji, Jiawei Sun, Renqi Chen, Wei Gao, Mani Srivastava

Mental health risk is a critical global public health challenge, necessitating innovative and reliable assessment methods. With the development of large language models (LLMs), they stand out to be a promising tool for explainable mental health care applications. Nevertheless, existing approaches predominantly rely on subjective textual mental records, which can be distorted by inherent mental uncertainties, leading to inconsistent and unreliable predictions. To address these limitations, this paper introduces ProMind-LLM. We investigate an innovative approach integrating objective behavior data as complementary information alongside subjective mental records for robust mental health risk assessment. Specifically, ProMind-LLM incorporates a comprehensive pipeline that includes domain-specific pretraining to tailor the LLM for mental health contexts, a self-refine mechanism to optimize the processing of numerical behavioral data, and causal chain-of-thought reasoning to enhance the reliability and interpretability of its predictions. Evaluations of two real-world datasets, PMData and Globem, demonstrate the effectiveness of our proposed methods, achieving substantial improvements over general LLMs. We anticipate that ProMind-LLM will pave the way for more dependable, interpretable, and scalable mental health case solutions.

Feng Luo, Rui Yang, Hao Sun, Chunyuan Deng, Jiarui Yao, Jingyan Shen, Huan Zhang, Hanjie Chen

Understanding human preferences is crucial for improving foundation models and building personalized AI systems. However, preferences are inherently diverse and complex, making it difficult for traditional reward models to capture their full range. While fine-grained preference data can help, collecting it is expensive and hard to scale. In this paper, we introduce Decomposed Reward Models (DRMs), a novel approach that extracts diverse human preferences from binary comparisons without requiring fine-grained annotations. Our key insight is to represent human preferences as vectors and analyze them using Principal Component Analysis (PCA). By constructing a dataset of embedding differences between preferred and rejected responses, DRMs identify orthogonal basis vectors that capture distinct aspects of preference. These decomposed rewards can be flexibly combined to align with different user needs, offering an interpretable and scalable alternative to traditional reward models. We demonstrate that DRMs effectively extract meaningful preference dimensions (e.g., helpfulness, safety, humor) and adapt to new users without additional training. Our results highlight DRMs as a powerful framework for personalized and interpretable LLM alignment.

Sirui Hong, Yizhang Lin, Bang Liu, Bangbang Liu, Binhao Wu, Ceyao Zhang, Danyang Li, Jiaqi Chen, Jiayi Zhang, Jinlin Wang 等

Large Language Model (LLM)-based agents have excelled in various domains but face significant challenges when applied to data science workflows due to their complex, multi-stage nature. Current LLM-based agents struggle with non-linear relationships, recursive dependencies, implicit data- and logic-dependent reasoning, and managing extensive context. In this paper, we introduce Data Interpreter, an LLM-based agent that addresses these challenges through hierarchical graph-based modeling to represent the complexity and a progressive strategy for step-by-step verification, refinement, and consistent context management. Extensive experiments confirm the effectiveness of Data Interpreter. On InfiAgent-DABench, it boosts performance by 25% (from 75.9% to 94.9%), and on machine learning and open-ended tasks, it lifts accuracy from 88% to 95% and from 60% to 97%, respectively. Moreover, our method surpasses state-of-the-art baselines by 26% on the MATH dataset. We will release the code upon publication.

Matthew King-Hang Ma, Xie Chenwei, Wenbo Wang, William Shiyuan Wang

Homonymy can easily raise lexical ambiguity due to the misunderstanding of its multiple senses. Correct recognition of homonym sense greatly relies on its surrounding context. This ambiguous nature makes homonyms an appropriate testbed for examining the contextualization capability of pre-trained (PLM) and large language models (LLMs). Considering the impact of part of speech (POS) on homonym disambiguation and the prevalence of English-focused studies in word embedding research, this study extends to Chinese and provides a comprehensive layer-wise analysis of homonym representations in both languages, spanning same and different POS categories, across four families of PLMs/LLMs (BERT, GPT-2, Llama 3, Qwen 2.5). Through the creation of a synthetic dataset and computation of disambiguation score (D-Score), we found that: (1) no universal layer depth excels in differentiating homonym representations; (2) bidirectional models produce better contextualized homonym representations compared to much larger autoregressive models; (3) most importantly, POS affects homonym representations in models in ways that differ from human research findings. The individual differences between LLMs uncovered in our study challenge the simplistic understanding of their inner workings. This reveals a compelling research frontier: conducting controlled experiments with purposefully manipulated inputs to enhance the interpretability of LLMs. We have made our dataset and codes available publicly at https://github.com/neurothew/exploring-homonym-rep-in-llm.

Ziyu Guo, Renrui Zhang, Hao Chen, Jialin Gao, Dongzhi Jiang, Jiaze Wang, Pheng-Ann Heng

The rapid advancement of Large Multi-modal Models (LMMs) has enabled their application in scientific problem-solving, yet their fine-grained capabilities remain under-explored. In this paper, we introduce SciVerse, a multi-modal scientific evaluation benchmark to thoroughly assess LMMs across 5,735 test instances in five distinct versions. We aim to investigate three key dimensions of LMMs: scientific knowledge comprehension, multi-modal content interpretation, and Chain-of-Thought (CoT) reasoning. To unveil whether LMMs possess sufficient scientific expertise, we first transform each problem into three versions containing different levels of knowledge required for solving, i.e., Knowledge-free, -lite, and -rich. Then, to explore how LMMs interpret multi-modal scientific content, we annotate another two versions, i.e., Vision-rich and -only, marking more question information from texts to diagrams. Comparing the results of different versions, SciVerse systematically examines the professional knowledge stock and visual perception skills of LMMs in scientific domains. In addition, to rigorously assess CoT reasoning, we propose a new scientific CoT evaluation strategy, conducting a step-wise assessment on knowledge and logical errors in model outputs. Our extensive evaluation of different LMMs on SciVerse reveals critical limitations in their scientific proficiency and provides new insights into future developments. Project page: https://sciverse-cuhk.github.io

Ahmed Masry, Mohammed Saidul Islam, Mahir Ahmed, Aayush Bajaj, Firoz Kabir, Aaryaman Kartha, Md Tahmid Rahman Laskar, Mizanur Rahman, Shadikur Rahman, Mehrad Shahmohammadi 等

Charts are ubiquitous, as people often use them to analyze data, answer questions, and discover critical insights. However, performing complex analytical tasks with charts requires significant perceptual and cognitive effort. Chart Question Answering (CQA) systems automate this process by enabling models to interpret and reason with visual representations of data. However, existing benchmarks like ChartQA lack real-world diversity and have recently shown performance saturation with modern large vision-language models (LVLMs). To address these limitations, we introduce ChartQAPro, a new benchmark that includes 1,341 charts from 99 diverse sources, spanning various chart types—including infographics and dashboards—and featuring 1,948 questions in various types, such as multiple-choice, conversational, hypothetical, and unanswerable questions, to better reflect real-world challenges. Our evaluations with 21 models show a substantial performance drop for LVLMs on ChartQAPro; e.g., Claude Sonnet 3.5 scores 90.5% on ChartQA but only 55.81% on ChartQAPro, underscoring the complexity of chart reasoning. We complement our findings with detailed error analyses and ablation studies, identifying key challenges and opportunities for advancing LVLMs in chart understanding and reasoning. We release ChartQAPro at https://github.com/vis-nlp/ChartQAPro.

Yuyou Zhang, Miao Li, William Han, Yihang Yao, Zhepeng Cen, Ding Zhao

Large Language Models (LLMs) are vulnerable to jailbreak attacks that exploit weaknesses in traditional safety alignment, which often relies on rigid refusal heuristics or representation engineering to block harmful outputs. While they are effective for direct adversarial attacks, they fall short of broader safety challenges requiring nuanced, context-aware decision-making. To address this, we propose Reasoning-enhanced Fine-Tuning for interpretable LLM Safety (RATIONAL), a novel framework that trains models to engage in explicit safe reasoning before response. Fine-tuned models leverage the extensive pretraining knowledge in self-generated reasoning to bootstrap their own safety through structured reasoning, internalizing context-sensitive decision-making. Our findings suggest that safety extends beyond refusal, requiring context awareness for more robust, interpretable, and adaptive responses. Reasoning is not only a core capability of LLMs but also a fundamental mechanism for LLM safety. RATIONAL employs reasoning-enhanced fine-tuning, allowing it to reject harmful prompts while providing meaningful and context-aware responses in complex scenarios.

Yilun Zhao, Chengye Wang, Chuhan Li, Arman Cohan

This paper introduces MISS-QA, the first benchmark specifically designed to evaluate the ability of models to interpret schematic diagrams within scientific literature. MISS-QA comprises 3,000 expert-annotated examples over 983 scientific papers. In this benchmark, models are tasked with interpreting schematic diagrams that illustrate research overviews and answering corresponding information-seeking questions based on the broader context of the paper. To ensure reliable and consistent evaluation, we propose an automated evaluating protocol powered by open-source LLMs trained on human-scored data. We assess the performance of 18 frontier multimodal foundation models, including o1, Claude-3.5, Llama-3.2-Vision, and Qwen2-VL. We reveal a significant performance gap between these models and human experts on MISS-QA. Our analysis of model performance on unanswerable questions and our detailed error analysis further highlight the strengths and limitations of current models, offering key insights to enhance models in comprehending multimodal scientific literature.

Denitsa Saynova, Lovisa Hagström, Moa Johansson, Richard Johansson, Marco Kuhlmann

Language models (LMs) can make a correct prediction based on many possible signals in a prompt, not all corresponding to recall of factual associations. However, current interpretations of LMs fail to take this into account. For example, given the query “Astrid Lindgren was born in” with the corresponding completion “Sweden”, no difference is made between whether the prediction was based on knowing where the author was born or assuming that a person with a Swedish-sounding name was born in Sweden. In this paper, we present a model-specific recipe - PrISM - for constructing datasets with examples of four different prediction scenarios: generic language modeling, guesswork, heuristics recall and exact fact recall. We apply two popular interpretability methods to the scenarios: causal tracing (CT) and information flow analysis. We find that both yield distinct results for each scenario. Results for exact fact recall and generic language modeling scenarios confirm previous conclusions about the importance of mid-range MLP sublayers for fact recall, while results for guesswork and heuristics indicate a critical role of late last token position MLP sublayers. In summary, we contribute resources for a more extensive and granular study of fact completion in LMs, together with analyses that provide a more nuanced understanding of how LMs process fact-related queries.

Iqra Zahid, Youcheng Sun, Riza Batista-Navarro

Authorship Attribution (AA) seeks to identify the author of a given text, yet existing methods often struggle with trustworthiness and interpretability, particularly across different domains, languages, and stylistic variations. These challenges arise from the absence of uncertainty quantification and the inability of current models to adapt to diverse authorship tasks. To address these limitations, we introduce BEDAA, a Bayesian-Enhanced DeBERTa framework that integrates Bayesian reasoning with transformer-based language models to enable uncertainty-aware and interpretable authorship attribution. BEDAA achieves up to 19.69% improvement in F1-score across multiple authorship attribution tasks, including binary, multiclass, and dynamic authorship detection. By incorporating confidence ranking, uncertainty decomposition, and probabilistic reasoning, BEDAA improves robustness while offering transparent decision-making processes. Furthermore, BEDAA extends beyond traditional AA by demonstrating its effectiveness in human vs. machine-generated text classification, code authorship detection, and cross-lingual attribution. These advances establish BEDAA as a generalised, interpretable, and adaptable framework for modern authorship attribution challenges.

Alireza Salemi, Julian Killingback, Hamed Zamani

Evaluating personalized text generated by large language models (LLMs) is challenging, as only the LLM user, i.e. prompt author, can reliably assess the output, but re-engaging the same individuals across studies is infeasible. This paper addresses the challenge of evaluating personalized text generation by introducing ExPerT, an explainable reference-based evaluation framework. ExPerT leverages an LLM to extract atomic aspects and their evidences from the generated and reference texts, match the aspects, and evaluate their alignment based on content and writing style—two key attributes in personalized text generation. Additionally, ExPerT generates detailed, fine-grained explanations for every step of the evaluation process, enhancing transparency and interpretability. Our experiments demonstrate that ExPerT achieves a 7.2% relative improvement in alignment with human judgments compared to the state-of-the-art text generation evaluation methods. Furthermore, human evaluators rated the usability of ExPerT’s explanations at 4.7 out of 5, highlighting its effectiveness in making evaluation decisions more interpretable.

Elisa Sanchez-Bayona, Rodrigo Agerri

This paper presents a comprehensive evaluation of the capabilities of Large Language Models (LLMs) in metaphor interpretation across multiple datasets, tasks, and prompt configurations. Although metaphor processing has gained significant attention in Natural Language Processing (NLP), previous research has been limited to single-dataset evaluations and specific task settings, often using artificially constructed data through lexical replacement. We address these limitations by conducting extensive experiments using diverse publicly available datasets with inference and metaphor annotations, focusing on Natural Language Inference (NLI) and Question Answering (QA) tasks. The results indicate that LLMs’ performance is more influenced by features like lexical overlap and sentence length than by metaphorical content, demonstrating that any alleged emergent abilities of LLMs to understand metaphorical language are the result of a combination of surface-level features, in-context learning, and linguistic knowledge. This work provides critical insights into the current capabilities and limitations of LLMs in processing figurative language, highlighting the need for more realistic evaluation frameworks in metaphor interpretation tasks. Data and code publicly available: https://github.com/elisanchez-beep/metaphorLLM

Jihao Shi, Xiao Ding, Kai Xiong, Hengwei Zhao, Bing Qin, Ting Liu

Entailment trees are essential for enhancing interpretability and transparency in tasks like question answering and natural language understanding. However, existing approaches often lack logical consistency, as they rely on static reward structures or ignore the intricate dependencies within multi-step reasoning. To address these limitations, we propose a method that integrates natural logic principles into reinforcement learning, enabling dynamic reward computation to guide entailment tree generation. Our approach ensures logical consistency across reasoning steps while improving interpretability and generalization. Experiments on EntailmentBank demonstrate significant improvements over state-of-the-art methods, highlighting the effectiveness of natural logic in structured reasoning.

Martina Miliani, Serena Auriemma, Alessandro Bondielli, Emmanuele Chersoni, Lucia C. Passaro, Irene Sucameli, Alessandro Lenci

Large Language Models (LLMs) are increasingly used in tasks requiring interpretive and inferential accuracy. In this paper, we introduce ExpliCa, a new dataset for evaluating LLMs in explicit causal reasoning. ExpliCa uniquely integrates both causal and temporal relations presented in different linguistic orders and explicitly expressed by linguistic connectives. The dataset is enriched with crowdsourced human acceptability ratings. We tested LLMs on ExpliCa through prompting and perplexity-based metrics. We assessed seven commercial and open-source LLMs, revealing that even top models struggle to reach 0.80 accuracy. Interestingly, models tend to confound temporal relations with causal ones, and their performance is also strongly influenced by the linguistic order of the events. Finally, perplexity-based scores and prompting performance are differently affected by model size.

Bruno Puri, Aakriti Jain, Elena Golimblevskaia, Patrick Kahardipraja, Thomas Wiegand, Wojciech Samek, Sebastian Lapuschkin

Recent advances in mechanistic interpretability have highlighted the potential of automating interpretability pipelines in analyzing the latent representations within LLMs. While this may enhance our understanding of internal mechanisms, the field lacks standardized evaluation methods for assessing the validity of discovered features. We attempt to bridge this gap by introducing **FADE**: Feature Alignment to Description Evaluation, a scalable model-agnostic framework for automatically evaluating feature-to-description alignment. **FADE** evaluates alignment across four key metrics – *Clarity, Responsiveness, Purity, and Faithfulness* – and systematically quantifies the causes of the misalignment between features and their descriptions. We apply **FADE** to analyze existing open-source feature descriptions and assess key components of automated interpretability pipelines, aiming to enhance the quality of descriptions. Our findings highlight fundamental challenges in generating feature descriptions, particularly for SAEs compared to MLP neurons, providing insights into the limitations and future directions of automated interpretability. We release **FADE** as an open-source package at: [github.com/brunibrun/FADE](https://github.com/brunibrun/FADE).

Chenxi Wang, Tianle Gu, Zhongyu Wei, Lang Gao, Zirui Song, Xiuying Chen

Human readers can efficiently comprehend scrambled words, a phenomenon known as \textit{Typoglycemia}, primarily by relying on word form; if word form alone is insufficient, they further utilize contextual cues for interpretation. While advanced large language models (LLMs) exhibit similar abilities, the underlying mechanisms remain unclear. To investigate this, we conduct controlled experiments to analyze the roles of word form and contextual information in semantic reconstruction and examine LLM attention patterns. Specifically, we first propose \textit{SemRecScore}, a reliable metric to quantify the degree of semantic reconstruction, and validate its effectiveness. Using this metric, we study how word form and contextual information influence LLMs’ semantic reconstruction ability, identifying \textit{word form as the core factor} in this process. Furthermore, we analyze \textit{how} LLMs utilize word form and find that they rely on specialized attention heads to extract and process word form information, with this mechanism remaining stable across varying levels of word scrambling. This distinction between LLMs’ fixed attention patterns primarily focused on word form and human readers’ adaptive strategy in balancing word form and contextual information provides insights into enhancing LLM performance by incorporating human-like, context-aware mechanisms. Code is available on: https://github.com/Aurora-cx/TypoLLM.

Irina Saparina, Mirella Lapata

Handling ambiguity and underspecification is an important challenge in natural language interfaces, particularly for tasks like text-to-SQL semantic parsing. We propose a modular approach that resolves ambiguity using natural language interpretations before mapping these to logical forms (e.g., SQL queries). Although LLMs excel at parsing unambiguous utterances, they show strong biases for ambiguous ones, typically predicting only preferred interpretations. We constructively exploit this bias to generate an initial set of preferred disambiguations and then apply a specialized infilling model to identify and generate missing interpretations. To train the infilling model, we introduce an annotation method that uses SQL execution to validate different meanings. Our approach improves interpretation coverage and generalizes across datasets with different annotation styles, database structures, and ambiguity types.

Masaki Sakata, Benjamin Heinzerling, Sho Yokoi, Takumi Ito, Kentaro Inui

We analyze the extent to which internal representations of language models (LMs) identify and distinguish mentions of named entities, focusing on the many-to-many correspondence between entities and their mentions.We first formulate two problems of entity mentions — ambiguity and variability — and propose a framework analogous to clustering quality metrics. Specifically, we quantify through cluster analysis of LM internal representations the extent to which mentions of the same entity cluster together and mentions of different entities remain separated.Our experiments examine five Transformer-based autoregressive models, showing that they effectively identify and distinguish entities with metrics analogous to precision and recall ranging from 0.66 to 0.9.Further analysis reveals that entity-related information is compactly represented in a low-dimensional linear subspace at early LM layers.Additionally, we clarify how the characteristics of entity representations influence word prediction performance.These findings are interpreted through the lens of isomorphism between LM representations and entity-centric knowledge structures in the real world, providing insights into how LMs internally organize and use entity information.

Dana Ramati, Daniela Gottesman, Mor Geva

Continuous prompts, or “soft prompts”, are a widely-adopted parameter-efficient tuning strategy for large language models, but are often less favorable due to their opaque nature. Prior attempts to interpret continuous prompts relied on projecting individual prompt tokens onto the vocabulary space. However, this approach is problematic as performant prompts can yield arbitrary or contradictory text, and it individually interprets each prompt token. In this work, we propose a new approach to interpret continuous prompts that elicits textual descriptions from their representations during model inference. Using a Patchscopes variant (Ghandeharioun et al., 2024) called InSPEcT over various tasks, we show our method often yields accurate task descriptions which become more faithful as task performance increases. Moreover, an elaborated version of InSPEcT reveals biased features in continuous prompts, whose presence correlates with biased model predictions. Providing an effective interpretability solution, InSPEcT can be leveraged to debug unwanted properties in continuous prompts and inform developers on ways to mitigate them.