Peer Revenue Play credits. Real papers.

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computation and Language

cs.CL Watch Computation and Language

Friday 23 January

Computation and Language

Relating Word Embedding Gender Biases to Gender Gaps: A Cross-Cultural Analysis

Scott Friedman, Sonja Schmer-Galunder, Anthony Chen et al. (+1 more)

Modern models for common NLP tasks often employ machine learning techniques and train on journalistic, social media, or other culturally-derived text. These have recently been scrutinized for racial and gender biases, rooting from inherent bias in their training text. These biases are often sub-optimal and recent work poses methods to rectify them; however, these biases may shed light on actual racial or gender gaps in the culture(s) that produced the training text, thereby helping us understand cultural context through big data. This paper presents an approach for quantifying gender bias in word embeddings, and then using them to characterize statistical gender gaps in education, politics, economics, and health. We validate these metrics on 2018 Twitter data spanning 51 U.S. regions and 99 countries. We correlate state and country word embedding biases with 18 international and 5 U.S.-based statistical gender gaps, characterizing regularities and predictive strength.
Read the full abstract

All authors Scott Friedman, Sonja Schmer-Galunder, Anthony Chen, Jeffrey Rye

Computation and Language

Reasoning Beyond Literal: Cross-style Multimodal Reasoning for Figurative Language Understanding

Seyyed Saeid Cheshmi, Hahnemann Ortiz, James Mooney et al. (+1 more)

Vision-language models (VLMs) have demonstrated strong reasoning abilities in literal multimodal tasks such as visual mathematics and science question answering. However, figurative language, such as sarcasm, humor, and metaphor, remains a significant challenge, as it conveys intent and emotion through subtle incongruities between expressed and intended meanings. In multimodal settings, accompanying images can amplify or invert textual meaning, demanding models that reason across modalities and account for subjectivity. We propose a three-step framework for developing efficient multimodal reasoning models that can (i) interpret multimodal figurative language, (ii) provide transparent reasoning traces, and (iii) generalize across multiple figurative styles. Experiments across four styles show that (1) incorporating reasoning traces substantially improves multimodal figurative understanding, (2) reasoning learned in one style can transfer to others, especially between related styles like sarcasm and humor, and (3) training jointly across styles yields a generalized reasoning VLM that outperforms much larger open- and closed-source models. Our findings show that lightweight VLMs with verifiable reasoning achieve robust cross-style generalization while providing inspectable reasoning traces for multimodal tasks. The code and implementation are available at https://github.com/scheshmi/CrossStyle-MMR.
Read the full abstract

All authors Seyyed Saeid Cheshmi, Hahnemann Ortiz, James Mooney, Dongyeop Kang

Computation and Language

Systematicity between Forms and Meanings across Languages Supports Efficient Communication

Doreen Osmelak, Yang Xu, Michael Hahn et al. (+1 more)

Languages vary widely in how meanings map to word forms. These mappings have been found to support efficient communication; however, this theory does not account for systematic relations within word forms. We examine how a restricted set of grammatical meanings (e.g. person, number) are expressed on verbs and pronouns across typologically diverse languages. Consistent with prior work, we find that verb and pronoun forms are shaped by competing communicative pressures for simplicity (minimizing the inventory of grammatical distinctions) and accuracy (enabling recovery of intended meanings). Crucially, our proposed model uses a novel measure of complexity (inverse of simplicity) based on the learnability of meaning-to-form mappings. This innovation captures fine-grained regularities in linguistic form, allowing better discrimination between attested and unattested systems, and establishes a new connection from efficient communication theory to systematicity in natural language.
Read the full abstract

All authors Doreen Osmelak, Yang Xu, Michael Hahn, Kate McCurdy

Computation and Language

Beyond Factual QA: Mentorship-Oriented Question Answering over Long-Form Multilingual Content

Parth Bhalerao, Ruiwen Guan, Diola Dsouza et al. (+1 more)

Question answering systems are typically evaluated on factual correctness, yet many real-world applications-such as education and career guidance-require mentorship: responses that provide reflection and guidance. Existing QA benchmarks rarely capture this distinction, particularly in multilingual and long-form settings. We introduce MentorQA, the first multilingual dataset and evaluation framework for mentorship-focused question answering from long-form videos, comprising nearly 9,000 QA pairs from 180 hours of content across four languages. We define mentorship-focused evaluation dimensions that go beyond factual accuracy, capturing clarity, alignment, and learning value. Using MentorQA, we compare Single-Agent, Dual-Agent, RAG, and Multi-Agent QA architectures under controlled conditions. Multi-Agent pipelines consistently produce higher-quality mentorship responses, with especially strong gains for complex topics and lower-resource languages. We further analyze the reliability of automated LLM-based evaluation, observing substantial variation in alignment with human judgments. Overall, this work establishes mentorship-focused QA as a distinct research problem and provides a multilingual benchmark for studying agentic architectures and evaluation design in educational AI. The dataset and evaluation framework are released at https://github.com/AIM-SCU/MentorQA.
Read the full abstract

All authors Parth Bhalerao, Ruiwen Guan, Diola Dsouza, Oana Ignat

Computation and Language

Who Gets Which Message? Auditing Demographic Bias in LLM-Generated Targeted Text

Tunazzina Islam

Large language models (LLMs) are increasingly capable of generating personalized, persuasive text at scale, raising new questions about bias and fairness in automated communication. This paper presents the first systematic analysis of how LLMs behave when tasked with demographic-conditioned targeted messaging. We introduce a controlled evaluation framework using three leading models: GPT-4o, Llama-3.3, and Mistral-Large-2.1, across two generation settings: Standalone Generation, which isolates intrinsic demographic effects, and Context-Rich Generation, which incorporates thematic and regional context to emulate realistic targeting. We evaluate generated messages along three dimensions: lexical content, language style, and persuasive framing. We instantiate this framework on climate communication and find consistent age- and gender-based asymmetries across models: male- and youth-targeted messages tend to emphasize more assertive and progressive framing, while female- and senior-targeted messages more often reflect warmth, care, and traditional themes. Contextual prompts systematically amplify these disparities, with persuasion scores being higher for male-targeted messages, while age-related differences vary across models. Our findings demonstrate how demographic stereotypes can surface and intensify in LLM-generated targeted communication, underscoring the need for bias-aware generation pipelines and transparent auditing frameworks that explicitly account for demographic conditioning in socially sensitive applications.
Read the full abstract

Computation and Language

Interpretability of the Intent Detection Problem: A New Approach

Eduardo Sanchez-Karhunen, Jose F. Quesada-Moreno and Miguel A. Gutiérrez-Naranjo

Intent detection, a fundamental text classification task, aims to identify and label the semantics of user queries, playing a vital role in numerous business applications. Despite the dominance of deep learning techniques in this field, the internal mechanisms enabling Recurrent Neural Networks (RNNs) to solve intent detection tasks are poorly understood. In this work, we apply dynamical systems theory to analyze how RNN architectures address this problem, using both the balanced SNIPS and the imbalanced ATIS datasets. By interpreting sentences as trajectories in the hidden state space, we first show that on the balanced SNIPS dataset, the network learns an ideal solution: the state space, constrained to a low-dimensional manifold, is partitioned into distinct clusters corresponding to each intent. The application of this framework to the imbalanced ATIS dataset then reveals how this ideal geometric solution is distorted by class imbalance, causing the clusters for low-frequency intents to degrade. Our framework decouples geometric separation from readout alignment, providing a novel, mechanistic explanation for real world performance disparities. These findings provide new insights into RNN dynamics, offering a geometric interpretation of how dataset properties directly shape a network's computational solution.
Read the full abstract

Computation and Language

Dynamic Role Assignment for Multi-Agent Debate

Miao Zhang, Junsik Kim, Siyuan Xiang et al. (+2 more)

Multi-agent large language model (LLM) and vision-language model (VLM) debate systems employ specialized roles for complex problem-solving, yet model specializations are not leveraged to decide which model should fill which role. We propose dynamic role assignment, a framework that runs a Meta-Debate to select suitable agents before the actual debate. The meta-debate has two stages: (1) proposal, where candidates provide role-tailored arguments, and (2) peer review, where proposals are scored with data and role-specific criteria to choose the best agent for each position. We evaluate our method on LLM problem solving benchmarks. Applied on top of existing debate systems, our approach consistently outperforms uniform assignments (filling all roles with the same model) by up to 74.8% and random assignments (assigning models to roles without considering their suitability) by up to 29.7%, depending on the task and the specific assignment. This work establishes a new paradigm for multi-agent system design, shifting from static agent deployment to dynamic and capability-aware selection.
Read the full abstract

All authors Miao Zhang, Junsik Kim, Siyuan Xiang, Jian Gao, Cheng Cao

Computation and Language · Cross-listed from cs.CV

Scaling medical imaging report generation with multimodal reinforcement learning

Qianchu Liu, Sheng Zhang, Guanghui Qin et al. (+10 more)

Frontier models have demonstrated remarkable capabilities in understanding and reasoning with natural-language text, but they still exhibit major competency gaps in multimodal understanding and reasoning especially in high-value verticals such as biomedicine. Medical imaging report generation is a prominent example. Supervised fine-tuning can substantially improve performance, but they are prone to overfitting to superficial boilerplate patterns. In this paper, we introduce Universal Report Generation (UniRG) as a general framework for medical imaging report generation. By leveraging reinforcement learning as a unifying mechanism to directly optimize for evaluation metrics designed for end applications, UniRG can significantly improve upon supervised fine-tuning and attain durable generalization across diverse institutions and clinical practices. We trained UniRG-CXR on publicly available chest X-ray (CXR) data and conducted a thorough evaluation in CXR report generation with rigorous evaluation scenarios. On the authoritative ReXrank benchmark, UniRG-CXR sets new overall SOTA, outperforming prior state of the art by a wide margin. We release our model at https://huggingface.co/microsoft/UniRG-CXR.
Read the full abstract

All authors Qianchu Liu, Sheng Zhang, Guanghui Qin, Yu Gu, Ying Jin, Sam Preston, Yanbo Xu, Sid Kiblawi, Wen-wai Yim, Timothy Ossowski, Tristan Naumann, Mu Wei, Hoifung Poon

Computation and Language

From Emotion to Expression: Theoretical Foundations and Resources for Fear Speech

Vigneshwaran Shankaran, Gabriella Lapesa and Claudia Wagner

Few forces rival fear in their ability to mobilize societies, distort communication, and reshape collective behavior. In computational linguistics, fear is primarily studied as an emotion, but not as a distinct form of speech. Fear speech content is widespread and growing, and often outperforms hate-speech content in reach and engagement because it appears "civiler" and evades moderation. Yet the computational study of fear speech remains fragmented and under-resourced. This can be understood by recognizing that fear speech is a phenomenon shaped by contributions from multiple disciplines. In this paper, we bridge cross-disciplinary perspectives by comparing theories of fear from Psychology, Political science, Communication science, and Linguistics. Building on this, we review existing definitions. We follow up with a survey of datasets from related research areas and propose a taxonomy that consolidates different dimensions of fear for studying fear speech. By reviewing current datasets and defining core concepts, our work offers both theoretical and practical guidance for creating datasets and advancing fear speech research.
Read the full abstract

Computation and Language · Cross-listed from cs.CY

Beyond Instrumental and Substitutive Paradigms: Introducing Machine Culture as an Emergent Phenomenon in Large Language Models

Yueqing Hu, Xinyang Peng, Yukun Zhao et al. (+3 more)

Recent scholarship typically characterizes Large Language Models (LLMs) through either an Instrumental Paradigm (viewing models as reflections of their developers' culture) or a Substitutive Paradigm (viewing models as bilingual proxies that switch cultural frames based on language). This study challenges these anthropomorphic frameworks by proposing Machine Culture as an emergent, distinct phenomenon. We employed a 2 (Model Origin: US vs. China) $\times$ 2 (Prompt Language: English vs. Chinese) factorial design across eight multimodal tasks, uniquely incorporating image generation and interpretation to extend analysis beyond textual boundaries. Results revealed inconsistencies with both dominant paradigms: Model origin did not predict cultural alignment, with US models frequently exhibiting “holistic” traits typically associated with East Asian data. Similarly, prompt language did not trigger stable cultural frame-switching; instead, we observed Cultural Reversal, where English prompts paradoxically elicited higher contextual attention than Chinese prompts. Crucially, we identified a novel phenomenon termed Service Persona Camouflage: Reinforcement Learning from Human Feedback (RLHF) collapsed cultural variance in affective tasks into a hyper-positive, zero-variance “helpful assistant” persona. We conclude that LLMs do not simulate human culture but exhibit an emergent Machine Culture – a probabilistic phenomenon shaped by superposition in high-dimensional space and mode collapse from safety alignment.
Read the full abstract

All authors Yueqing Hu, Xinyang Peng, Yukun Zhao, Lin Qiu, Ka-lai Hung, Kaiping Peng

Computation and Language · Cross-listed from cs.LG

The Mouth is Not the Brain: Bridging Energy-Based World Models and Language Generation

Junichiro Niimi

Large Language Models (LLMs) generate fluent text, yet whether they truly understand the world or merely produce plausible texts about it remains contested. We propose an architectural principle, the mouth is not the brain, that explicitly separates world models from language models. Our architecture comprises three components: a DBM that captures domain structure as an energy-based world model, an adapter that projects latent belief states into embedding space, and a frozen GPT-2 that provides linguistic competence without domain knowledge. We instantiate this framework in the consumer review domain using Amazon smartphone reviews. Experiments demonstrate that (1) world model conditioning achieves lower cross-entropy loss and higher semantic similarity than architectural baselines including direct projection and full fine-tuning, while qualitative analysis reveals that soft prompt conditioning resolves a trade-off that prompt-based approaches cannot: simple prompts lack expressiveness while detailed prompts cause output collapse in small LLMs; (2) the DBM's energy function distinguishes coherent from incoherent market configurations, assigning higher energy to implausible brand-price combinations; and (3) interventions on specific attributes propagate causally to generated text with intervened outputs exhibiting distributions statistically consistent with naturally occurring samples sharing the target configuration. These findings suggest that even small-scale language models can achieve consistent, controllable generation when connected to an appropriate world model, providing empirical support for separating linguistic competence from world understanding.
Read the full abstract

Computation and Language · Cross-listed from cs.CY

Do VLMs Have a Moral Backbone? A Study on the Fragile Morality of Vision-Language Models

Zhining Liu, Tianyi Wang, Xiao Lin et al. (+9 more)

Despite substantial efforts toward improving the moral alignment of Vision-Language Models (VLMs), it remains unclear whether their ethical judgments are stable in realistic settings. This work studies moral robustness in VLMs, defined as the ability to preserve moral judgments under textual and visual perturbations that do not alter the underlying moral context. We systematically probe VLMs with a diverse set of model-agnostic multimodal perturbations and find that their moral stances are highly fragile, frequently flipping under simple manipulations. Our analysis reveals systematic vulnerabilities across perturbation types, moral domains, and model scales, including a sycophancy trade-off where stronger instruction-following models are more susceptible to persuasion. We further show that lightweight inference-time interventions can partially restore moral stability. These results demonstrate that moral alignment alone is insufficient and that moral robustness is a necessary criterion for the responsible deployment of VLMs.
Read the full abstract

All authors Zhining Liu, Tianyi Wang, Xiao Lin, Penghao Ouyang, Gaotang Li, Ze Yang, Hui Liu, Sumit Keswani, Vishwa Pardeshi, Huijun Zhao, Wei Fan, Hanghang Tong

Computation and Language

Strategies for Span Labeling with Large Language Models

Danil Semin, Ondřej Dušek and Zdeněk Kasner

Large language models (LLMs) are increasingly used for text analysis tasks, such as named entity recognition or error detection. Unlike encoder-based models, however, generative architectures lack an explicit mechanism to refer to specific parts of their input. This leads to a variety of ad-hoc prompting strategies for span labeling, often with inconsistent results. In this paper, we categorize these strategies into three families: tagging the input text, indexing numerical positions of spans, and matching span content. To address the limitations of content matching, we introduce LogitMatch, a new constrained decoding method that forces the model's output to align with valid input spans. We evaluate all methods across four diverse tasks. We find that while tagging remains a robust baseline, LogitMatch improves upon competitive matching-based methods by eliminating span matching issues and outperforms other strategies in some setups.
Read the full abstract

Computation and Language

Information Representation Fairness in Long-Document Embeddings: The Peculiar Interaction of Positional and Language Bias

Elias Schuhmacher, Andrianos Michail, Juri Opitz et al. (+2 more)

To be discoverable in an embedding-based search process, each part of a document should be reflected in its embedding representation. To quantify any potential reflection biases, we introduce a permutation-based evaluation framework. With this, we observe that state-of-the-art embedding models exhibit systematic positional and language biases when documents are longer and consist of multiple segments. Specifically, early segments and segments in higher-resource languages like English are over-represented, while later segments and segments in lower-resource languages are marginalized. In our further analysis, we find that the positional bias stems from front-loaded attention distributions in pooling-token embeddings, where early tokens receive more attention. To mitigate this issue, we introduce an inference-time attention calibration method that redistributes attention more evenly across document positions, increasing discoverabiltiy of later segments. Our evaluation framework and attention calibration is available at https://github.com/impresso/fair-sentence-transformers
Read the full abstract

All authors Elias Schuhmacher, Andrianos Michail, Juri Opitz, Rico Sennrich, Simon Clematide

Computation and Language

LLM-Based Adversarial Persuasion Attacks on Fact-Checking Systems

João A. Leite, Olesya Razuvayevskaya, Kalina Bontcheva et al. (+1 more)

Automated fact-checking (AFC) systems are susceptible to adversarial attacks, enabling false claims to evade detection. Existing adversarial frameworks typically rely on injecting noise or altering semantics, yet no existing framework exploits the adversarial potential of persuasion techniques against AFC systems, which are widely used in disinformation campaigns to manipulate audiences. In this paper, we introduce a novel class of persuasive adversarial attacks on AFCs by employing an LLM to rephrase claims using persuasion techniques. Considering $15$ techniques grouped into $5$ categories, we study the effects of persuasion on both claim verification and evidence retrieval using a decoupled evaluation strategy. Experiments on the FEVER and FEVEROUS benchmarks show that persuasion attacks can substantially degrade both verification performance and evidence retrieval. Our analysis identifies persuasion techniques as a potent class of adversarial attacks, highlighting the need for more robust AFC systems.
Read the full abstract

All authors João A. Leite, Olesya Razuvayevskaya, Kalina Bontcheva, Carolina Scarton

Computation and Language · Cross-listed from cs.AI

Reasoning Promotes Robustness in Theory of Mind Tasks

Ian B. de Haan, Peter van der Putten and Max van Duijn

Large language models (LLMs) have recently shown strong performance on Theory of Mind (ToM) tests, prompting debate about the nature and true performance of the underlying capabilities. At the same time, reasoning-oriented LLMs trained via reinforcement learning with verifiable rewards (RLVR) have achieved notable improvements across a range of benchmarks. This paper examines the behavior of such reasoning models in ToM tasks, using novel adaptations of machine psychological experiments and results from established benchmarks. We observe that reasoning models consistently exhibit increased robustness to prompt variations and task perturbations. Our analysis indicates that the observed gains are more plausibly attributed to increased robustness in finding the correct solution, rather than to fundamentally new forms of ToM reasoning. We discuss the implications of this interpretation for evaluating social-cognitive behavior in LLMs.
Read the full abstract

Computation and Language · Cross-listed from cs.CV

ColorConceptBench: A Benchmark for Probabilistic Color-Concept Understanding in Text-to-Image Models

Chenxi Ruan, Yihan Hou, Yu Xiao et al. (+2 more)

Text-to-image (T2I) models have advanced considerably in generating high-quality images from textual descriptions. However, their ability to associate colors with concepts remains largely constrained to explicit color names or codes, while their capacity to handle implicit concepts, such as emotions and visual states, remains underexplored. To address this gap, we introduce ColorConceptBench, an expert-annotated benchmark that systematically evaluates color-concept associations through probabilistic color distributions. ColorConceptBench moves beyond explicit color specifications by examining how models interpret 1,281 implicit color concepts, grounded in 6,584 human annotations. Our evaluation of nine leading T2I models reveals that performance varies substantially across semantic categories, and models exhibit a significant lack of sensitivity to abstract semantics. These limitations persist even when applying classifier-free guidance scaling at inference time, suggesting that achieving human-like color understanding demands a shift in how models learn and represent implicit semantic meaning.
Read the full abstract

All authors Chenxi Ruan, Yihan Hou, Yu Xiao, Guosheng Hu, Wei Zeng

Computation and Language

Disentangling generalization and memorization in large language models using chess

Leonard S. Pleiss, Maximilian Schiffer and Robert K. von Weizsaecker

Large Language Models (LLMs) exhibit remarkable capabilities, yet it remains unclear to what extent these reflect sophisticated recall or genuine reasoning ability. We introduce chess as a controlled testbed aimed at disentangling these faculties. Leveraging the game's structure and scalable engine evaluations, we construct a taxonomy of positions varying in density of relevant priors – ranging from common states solvable by memorization to completely novel ones requiring generalization. Crucially, our approach achieves this distinction without requiring explicit knowledge of the models' training data. Applying this taxonomy, we combine a longitudinal analysis of the GPT lineage with a rigorous evaluation of contemporary models, including Claude Opus and Gemini. Our analysis reveals a steep gradient: performance consistently degrades as the density of relevant priors decreases. Notably, for tasks with few relevant priors, base model performance regresses to the random-play baseline, and all models propose illegal moves at sharply increasing rates, indicating that they fail to reliably apply the fixed rules of the game in unfamiliar positions. While newer models improve, progress slows significantly for tasks with sparse priors. Furthermore, while reasoning-augmented inference improves performance, its relative marginal benefit per token decreases in the absence of relevant priors. These results suggest limitations in systematic generalization, highlighting the need for mechanisms beyond scale to achieve robust performance when deprived of relevant priors.
Read the full abstract

Computation and Language

SoS: Analysis of Surface over Semantics in Multilingual Text-To-Image Generation

Carolin Holtermann, Florian Schneider and Anne Lauscher

Text-to-image (T2I) models are increasingly employed by users worldwide. However, prior research has pointed to the high sensitivity of T2I towards particular input languages - when faced with languages other than English (i.e., different surface forms of the same prompt), T2I models often produce culturally stereotypical depictions, prioritizing the surface over the prompt's semantics. Yet a comprehensive analysis of this behavior, which we dub Surface-over-Semantics (SoS), is missing. We present the first analysis of T2I models' SoS tendencies. To this end, we create a set of prompts covering 171 cultural identities, translated into 14 languages, and use it to prompt seven T2I models. To quantify SoS tendencies across models, languages, and cultures, we introduce a novel measure and analyze how the tendencies we identify manifest visually. We show that all but one model exhibit strong surface-level tendency in at least two languages, with this effect intensifying across the layers of T2I text encoders. Moreover, these surface tendencies frequently correlate with stereotypical visual depictions.
Read the full abstract

Computation and Language

Large Language Models as Automatic Annotators and Annotation Adjudicators for Fine-Grained Opinion Analysis

Gaurav Negi, MA Waskow, John McCrae et al. (+2 more)

Fine-grained opinion analysis of text provides a detailed understanding of expressed sentiments and their targets. Although this level of detail is valuable, annotating opinions in datasets for model training requires considerable human effort and substantial cost, especially across diverse domains and real-world applications. To address this shortage of domain-specific labelled datasets, we explore the feasibility of LLMs as automatic annotators for fine-grained opinion analysis. We use a declarative annotation pipeline, an approach that reduces the variability of manual prompt engineering when using LLMs to identify fine-grained opinion spans in text. We also present a dedicated methodology for an LLM to adjudicate multiple labels and produce final annotations, benchmarked against exact, flexible, and element-wise variants of a rule-based voting aggregator. We trial the pipeline with models of different sizes for the Aspect Sentiment Triplet Extraction (ASTE) and Aspect-Category-Opinion-Sentiment (ACOS) analysis tasks. Our results reveal a critical performance bifurcation: LLMs are reliable at the span level yet struggle to reproduce the relational structures that connect those spans faithfully. This suggests that LLMs are better positioned as high-fidelity annotation assistants and data augmentation tools to expand fine-grained opinion-annotated datasets, rather than replacing human annotators entirely.
Read the full abstract

All authors Gaurav Negi, MA Waskow, John McCrae, Omnia Zayed, Paul Buitelaar

Computation and Language

Persuasion Tokens for Editing Factual Knowledge in LLMs

Paul Youssef, Christin Seifert and Jörg Schlötterer

In-context knowledge editing (IKE) is a promising technique for updating Large Language Models (LLMs) with new information. However, IKE relies on lengthy, fact-specific demonstrations which are costly to create and consume significant context window space. In this paper, we introduce persuasion tokens (P-Tokens) – special tokens trained to replicate the effect of IKE demonstrations, enabling efficient knowledge editing without requiring fact-specific demonstrations. We evaluate P-Tokens across two editing datasets and three LLMs, demonstrating performance comparable to, and often exceeding, IKE. We further find that editing performance is robust to distractors with small negative effects to neighboring facts, and that increasing the number of P-Tokens improves performance. Our work addresses key limitations of IKE and provides a more practical and scalable alternative for editing LLMs.
Read the full abstract

Computation and Language

Do LLM hallucination detectors suffer from low-resource effect?

Debtanu Datta, Mohan Kishore Chilukuri, Yash Kumar et al. (+2 more)

LLMs, while outperforming humans in a wide range of tasks, can still fail in unanticipated ways. We focus on two pervasive failure modes: (i) hallucinations, where models produce incorrect information about the world, and (ii) the low-resource effect, where the models show impressive performance in high-resource languages like English but the performance degrades significantly in low-resource languages like Bengali. We study the intersection of these issues and ask: do hallucination detectors suffer from the low-resource effect? We conduct experiments on five tasks across three domains (factual recall, STEM, and Humanities). Experiments with four LLMs and three hallucination detectors reveal a curious finding: As expected, the task accuracies in low-resource languages experience large drops (compared to English). However, the drop in detectors' accuracy is often several times smaller than the drop in task accuracy. Our findings suggest that even in low-resource languages, the internal mechanisms of LLMs might encode signals about their uncertainty. Further, the detectors are robust within language (even for non-English) and in multilingual setups, but not in cross-lingual settings without in-language supervision.
Read the full abstract

All authors Debtanu Datta, Mohan Kishore Chilukuri, Yash Kumar, Saptarshi Ghosh, Muhammad Bilal Zafar

Computation and Language

Standardizing Longitudinal Radiology Report Evaluation via Large Language Model Annotation

Xinyi Wang, Grazziela Figueredo, Ruizhe Li et al. (+1 more)

Longitudinal information in radiology reports refers to the sequential tracking of findings across multiple examinations over time, which is crucial for monitoring disease progression and guiding clinical decisions. Many recent automated radiology report generation methods are designed to capture longitudinal information; however, validating their performance is challenging. There is no proper tool to consistently label temporal changes in both ground-truth and model-generated texts for meaningful comparisons. Large language models (LLMs) offer a promising annotation alternative, as they are capable of capturing nuanced linguistic patterns and semantic similarities without extensive manual intervention. They also adapt well to new contexts. In this study, we therefore propose an LLM-based pipeline to automatically annotate longitudinal information in radiology reports. The pipeline first identifies sentences containing relevant information and then extracts the progression of diseases. We evaluate and compare five mainstream LLMs on these two tasks using 500 manually annotated reports. Considering both efficiency and performance, Qwen2.5-32B was subsequently selected and used to annotate another 95,169 reports from the public MIMIC-CXR dataset. Our Qwen2.5-32B-annotated dataset provided us with a standardized benchmark for evaluating report generation models. Using this new benchmark, we assessed seven state-of-the-art report generation models. Our LLM-based annotation method outperforms existing annotation solutions, achieving 11.3% and 5.3% higher F1-scores for longitudinal information detection and disease tracking, respectively. The source code is available at https://github.com/wxinyi1996/Standardizing-Longitudinal-Chest-X-ray-Report-Evaluation-via-Large-Language-Model-Annotation.git.
Read the full abstract

All authors Xinyi Wang, Grazziela Figueredo, Ruizhe Li, Xin Chen

Computation and Language · Cross-listed from cs.SE

SWE-Pruner: Self-Adaptive Context Pruning for Coding Agents

Yuhang Wang, Yuling Shi, Mo Yang et al. (+7 more)

LLM agents have demonstrated remarkable capabilities in software development, but their performance is hampered by long interaction contexts, which incur high API costs and latency. While various context compression approaches such as LongLLMLingua have emerged to tackle this challenge, they typically rely on fixed metrics such as PPL, ignoring the task-specific nature of code understanding. As a result, they frequently disrupt syntactic and logical structure and fail to retain critical implementation details. In this paper, we propose SWE-Pruner, a self-adaptive context pruning framework tailored for coding agents. Drawing inspiration from how human programmers "selectively skim" source code during development and debugging, SWE-Pruner performs task-aware adaptive pruning for long contexts. Given the current task, the agent formulates an explicit goal (e.g., "focus on error handling") as a hint to guide the pruning targets. A lightweight neural skimmer (0.6B parameters) is trained to dynamically select relevant lines from the surrounding context given the goal. Evaluations across four benchmarks and multiple models validate SWE-Pruner's effectiveness in various scenarios, achieving 23-54% token reduction on agent tasks like SWE-Bench Verified while even improving success rates, and up to 14.84x compression on single-turn tasks like LongCodeQA with minimal performance impact.
Read the full abstract

All authors Yuhang Wang, Yuling Shi, Mo Yang, Rongrui Zhang, Shilin He, Heng Lian, Yuting Chen, Siyu Ye, Kai Cai, Xiaodong Gu

Computation and Language

Mitigating Bias in Automated Essay Scoring for ESL Learners via Contrastive Learning

Kevin Fan and Eric Yun

Automated Essay Scoring systems disproportionately penalize high-proficiency English as a Second Language (ESL) learners. We propose Contrastive Learning with Matched Essay Pairs (CL-MEP), a bi-directional alignment strategy. CL-MEP reduces this scoring bias by 39.9% while improving overall accuracy, successfully disentangling valid syntactic complexity from surface-level grammatical errors.
Read the full abstract

Computation and Language

Better Generalizing to Unseen Concepts: An Evaluation Framework and An LLM-Based Auto-Labeled Pipeline for Biomedical Concept Recognition

Shanshan Liu, Noriki Nishida, Fei Cheng et al. (+6 more)

Generalization to unseen concepts is a central challenge due to the scarcity of human annotations in Mention-agnostic Biomedical Concept Recognition (MA-BCR). This work makes two key contributions to systematically address this issue. First, we propose an evaluation framework built on hierarchical concept indices and novel metrics to measure generalization. Second, we explore LLM-based Auto-Labeled Data (ALD) as a scalable resource, creating a task-specific pipeline for its generation. Our research unequivocally shows that while LLM-generated ALD cannot fully substitute for manual annotations, it is a valuable resource for improving generalization, successfully providing models with the broader coverage and structural knowledge needed to approach recognizing unseen concepts. Code and datasets are available at https://github.com/bio-ie-tool/hi-ald.
Read the full abstract

All authors Shanshan Liu, Noriki Nishida, Fei Cheng, Narumi Tokunaga, Rumana Ferdous Munne, Yuki Yamagata, Kouji Kozaki, Takehito Utsuro, Yuji Matsumoto

Computation and Language

EMemBench: Interactive Benchmarking of Episodic Memory for VLM Agents

Xinze Li, Ziyue Zhu, Siyuan Liu et al. (+4 more)

We introduce EMemBench, a programmatic benchmark generator for evaluating long-term episodic memory of agents through interactive games. Rather than using a fixed set of questions, EMemBench generates questions from environment-grounded trajectories, covering both text-only and visual game environments. Each template computes verifiable ground truth from underlying game signals, with controlled answerability and balanced coverage over memory skills: single/multi-hop recall, induction, temporal, spatial, logical, and adversarial. We evaluate memory agents with strong LMs/VLMs as backbones, using in-context prompting as baselines. Across 15 text games and multiple visual seeds, results are far from saturated: induction and spatial reasoning are persistent bottlenecks, especially in visual settings. Persistent memory yields clear gains for open backbones on text games, but improvements are less consistent for VLM agents, suggesting that visually grounded episodic memory remains an open challenge. A human study further contextualizes the difficulty and interpretability of EMemBench.
Read the full abstract

All authors Xinze Li, Ziyue Zhu, Siyuan Liu, Yubo Ma, Yuhang Zang, Yixin Cao, Aixin Sun

Computation and Language

PLawBench: A Rubric-Based Benchmark for Evaluating LLMs in Real-World Legal Practice

Yuzhen Shi, Huanghai Liu, Yiran Hu et al. (+27 more)

As large language models (LLMs) are increasingly applied to legal domain-specific tasks, evaluating their ability to perform legal work in real-world settings has become essential. However, existing legal benchmarks rely on simplified and highly standardized tasks, failing to capture the ambiguity, complexity, and reasoning demands of real legal practice. Moreover, prior evaluations often adopt coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning. To address these limitations, we introduce PLawBench, a Practical Law Benchmark designed to evaluate LLMs in realistic legal practice scenarios. Grounded in real-world legal workflows, PLawBench models the core processes of legal practitioners through three task categories: public legal consultation, practical case analysis, and legal document generation. These tasks assess a model's ability to identify legal issues and key facts, perform structured legal reasoning, and generate legally coherent documents. PLawBench comprises 850 questions across 13 practical legal scenarios, with each question accompanied by expert-designed evaluation rubrics, resulting in approximately 12,500 rubric items for fine-grained assessment. Using an LLM-based evaluator aligned with human expert judgments, we evaluate 10 state-of-the-art LLMs. Experimental results show that none achieves strong performance on PLawBench, revealing substantial limitations in the fine-grained legal reasoning capabilities of current LLMs and highlighting important directions for future evaluation and development of legal LLMs. Data is available at: https://github.com/skylenage/PLawbench.
Read the full abstract

All authors Yuzhen Shi, Huanghai Liu, Yiran Hu, Gaojie Song, Xinran Xu, Yubo Ma, Tianyi Tang, Li Zhang, Qingjing Chen, Di Feng, Wenbo Lv, Weiheng Wu, Kexin Yang, Sen Yang, Wei Wang, Rongyao Shi, Yuanyang Qiu, Yuemeng Qi, Jingwen Zhang, Xiaoyu Sui, Yifan Chen, Yi Zhang, An Yang, Bowen Yu, Dayiheng Liu, Junyang Lin, Weixing Shen, Bing Zhao, Charles L.A. Clarke, Hu Wei

Computation and Language

Select or Project? Evaluating Lower-dimensional Vectors for LLM Training Data Explanations

Lukas Hinterleitner, Loris Schoenegger and Benjamin Roth

Gradient-based methods for instance-based explanation for large language models (LLMs) are hindered by the immense dimensionality of model gradients. In practice, influence estimation is restricted to a subset of model parameters to make computation tractable, but this subset is often chosen ad hoc and rarely justified by systematic evaluation. This paper investigates if it is better to create low-dimensional representations by selecting a small, architecturally informed subset of model components or by projecting the full gradients into a lower-dimensional space. Using a novel benchmark, we show that a greedily selected subset of components captures the information about training data influence needed for a retrieval task more effectively than either the full gradient or random projection. We further find that this approach is more computationally efficient than random projection, demonstrating that targeted component selection is a practical strategy for making instance-based explanations of large models more computationally feasible.
Read the full abstract

Computation and Language

Sycophancy Hides Linearly in the Attention Heads

Rifo Genadi, Munachiso Nwadike, Nurdaulet Mukhituly et al. (+3 more)

We find that correct-to-incorrect sycophancy signals are most linearly separable within multi-head attention activations. Motivated by the linear representation hypothesis, we train linear probes across the residual stream, multilayer perceptron (MLP), and attention layers to analyze where these signals emerge. Although separability appears in the residual stream and MLPs, steering using these probes is most effective in a sparse subset of middle-layer attention heads. Using TruthfulQA as the base dataset, we find that probes trained on it transfer effectively to other factual QA benchmarks. Furthermore, comparing our discovered direction to previously identified "truthful" directions reveals limited overlap, suggesting that factual accuracy, and deference resistance, arise from related but distinct mechanisms. Attention-pattern analysis further indicates that the influential heads attend disproportionately to expressions of user doubt, contributing to sycophantic shifts. Overall, these findings suggest that sycophancy can be mitigated through simple, targeted linear interventions that exploit the internal geometry of attention activations.
Read the full abstract

All authors Rifo Genadi, Munachiso Nwadike, Nurdaulet Mukhituly, Hilal Alquabeh, Tatsuya Hiraoka, Kentaro Inui