Peer Revenue Past citations are no guarantee of future returns

Papers

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

Computer Science Computation and Language

cs.CL Watch Computation and Language

Saturday 6 June

Computation and Language

GlobeAudio: A Multilingual Multicultural Benchmark for Naturalistic Evaluation of Large Audio-Language Models

Ryner Tan and Wenxuan Zhang

Large Audio-Language Models (LALMs) integrate audio perception and language understanding within a unified framework, enabling a wide range of real-world applications. Despite recent advances, evaluation for LALMs remains heavily underspecified relative to real-world requirements: most lack true linguistic and cultural authenticity, while others fail to capture acoustic realism. To bridge this gap, we propose GlobeAudio, a multilingual and multicultural benchmark designed to evaluate naturalistic audio understanding. GlobeAudio consists of 5,637 multiple-choice questions across six typologically diverse languages, expertly crafted by native speakers grounded on naturally occurring audio. In order to do well, models must possess higher-level auditory reasoning skills and culturally grounded interpretation. We systematically evaluate representative closed-source and open-source LALMs, as well as cascaded ASR-LLM pipelines. Our experiments reveal substantial performance gaps under natural acoustic conditions, particularly for open-source models and low-resource languages. These findings highlight critical limitations of current LALMs and underscore the importance of naturalistic audio evaluation for future audio-language systems. GlobeAudio can be found at https://huggingface.co/datasets/iNLP-Lab/GlobeAudio .
Read the full abstract

Computation and Language

TextEconomizer: Enhancing Lossy Text Compression with Denoising Transformers and Entropy Coding

Mahbub E Sobhani, Anika Tasnim Rodela, Chowdhury Mofizur Rahman et al. (+2 more)

Lossy text compression reduces data size while preserving core meaning, making it well-suited for summarization, automated analysis, and digital archives. Despite the dominance of transformer-based models in language modeling, integrating context vectors and entropy coding into Sequence-to-Sequence (Seq2Seq) generation remains underexplored. A key challenge lies in identifying the most informative context vectors from encoder output and incorporating entropy coding to enhance storage efficiency while maintaining high-quality outputs, even under noisy text. We introduce TextEconomizer, an encoder-decoder framework paired with a transformer neural network that reduces variable-sized inputs by 50% to 80% without prior knowledge of dataset dimensions. Our model achieves competitive compression ratios via entropy coding while delivering near-perfect text quality, assessed by BLEU, ROUGE, METEOR, and semantic similarity scores. TextEconomizer operates with approximately 153x fewer parameters than comparable models, achieving a 5.39x compression ratio without sacrificing semantic quality. We also evaluate an LSTM-based autoencoder achieving a state-of-the-art 67x compression ratio with 196x fewer parameters, and LLaMAFormer, a modified transformer with 263x fewer parameters than ICAE while maintaining competitive text quality. TextEconomizer significantly surpasses existing transformer-based models in balancing memory efficiency and high-fidelity outputs, marking a breakthrough in lossy compression with optimal space utilization.
Read the full abstract

All authors Mahbub E Sobhani, Anika Tasnim Rodela, Chowdhury Mofizur Rahman, Dewan Md. Farid, Swakkhar Shatabda

Computation and Language · Cross-listed from cs.RO

CLASP: Language-Driven Robot Skill Selection and Composition using Task-Parameterized Learning

Markus Knauer, Valentin Gieraths, Tai Mai et al. (+4 more)

Enabling robots to understand and execute tasks from natural language commands while maintaining data efficiency remains challenging. Foundation models such as vision-language-action (VLA) and vision-language models (VLMs) provide intuitive interaction channels but require extensive data; task-parameterized imitation learning achieves data efficiency but lacks natural language grounding. This work bridges this gap through a modular architecture combining task-parameterized kernelized movement primitives (TP-KMPs) with pretrained VLMs. During learning, skills are acquired from 2 to 5 kinesthetic demonstrations, and the VLM generates skill schemas describing each skill's parameters and preconditions. During execution, the VLM interprets commands to select skills, reason about parameter bindings, and create novel behaviors through covariance-weighted composition. When no skill or composition suffices, the system identifies capability gaps and requests targeted demonstrations, all without fine-tuning. Validation on a 7-DoF manipulator shows success rates of 73.3%-100% in scenarios requiring skill selection, composition, and active learning.
Read the full abstract

All authors Markus Knauer, Valentin Gieraths, Tai Mai, Samuel Bustamante, Alin Albu-Schäffer, Freek Stulp, João Silvério

Computation and Language

Constrained Paraphrase Consistency for LLM Hallucination Detection

Shanshan Lin, Dongsheng Hong, Sibo Ju et al. (+3 more)

Large language models (LLMs) can generate factually inconsistent claims, motivating accurate and scalable hallucination detectors. Prior work largely enlarges training sets via synthesis or new annotations, introducing increasing cost and potential bias while underusing the consistency implied by semantically equivalent paraphrases. We propose Consistency-Constrained Hallucination Detector (CCHD), which formulates training as a constrained optimization problem. The standard cross-entropy on original document-claim pairs is complemented by (i) paraphrase-consistency constraints bounding divergence across paraphrased views, and (ii) label-preservation constraints tying paraphrases to ground truth. We solve the problem by gradient descent-ascent over model parameters and per-view Lagrange multipliers, adding only a few scalar dual variables and no inference-time overhead. With DeBERTa and Flan-T5 backbones, CCHD consistently outperforms strong baselines (FactCG, MiniCheck, and AlignScore) on standard factuality benchmarks, demonstrating its superiority on hallucination detection.
Read the full abstract

All authors Shanshan Lin, Dongsheng Hong, Sibo Ju, Chao Chen, Xi Zhang, Xiangwen Liao

Computation and Language

Cross Paraphrastic Invariance Learning for Hallucination Detection

Shanshan Lin, Dongsheng Hong, Sibo Ju et al. (+3 more)

Large language models (LLMs) frequently generate hallucinations, which are unsupported by a source document. To avoid costly LLM-as-evaluator pipelines and the heavy annotation demands of existing classifiers, we propose CPIL (Cross Paraphrastic Invariance Learning), a two-stage Siamese framework that maximizes the utility of existing labeled data. Concretely, CPIL constructs informative training pairs by: (i) generating paraphrastic views of each document-claim example as positives, and explicitly aligning their representations to enforce invariance to surface form; and (ii) mining same-document, opposite-label pairs as hard negatives to sharpen document-sensitive decision boundaries. Then CPIL conduct a two-stage model training: Stage 1 performs contrastive pretraining to learn a paraphrase-invariant, grounding-aware embedding space; and Stage 2 attaches a lightweight classifier for binary groundedness. On the LLM-AggreFact benchmark (11 tasks), CPIL surpasses strong baselines concerning F1 scores with only ~1% labeled data, showing its prediction superiority and label efficiency.
Read the full abstract

All authors Shanshan Lin, Dongsheng Hong, Sibo Ju, Chao Chen, Sihong Xie, Xiangwen Liao

Computation and Language

When Languages Disagree: Self-Evolving Multilingual LLM Judges

Xiyan Fu and Wei Lu

Multilingual LLM-as-a-judge is widely used to evaluate model outputs across languages, but suffers from cross-lingual inconsistency (Fu and Liu, 2025). Existing methods typically treat this inconsistency as noise and mitigate it through voting or aggregation. In this work, we instead show that multilingual inconsistency can provide complementary evaluation signals. Our oracle analysis finds that sampling judgments across languages yields a higher performance upper bound than single-language judging, indicating that different languages potentially include complementary judgments. Motivated by this finding, we propose SEMJ, a self-evolving multilingual judge that leverages cross-lingual inconsistency for iterative refinement. SEMJ constructs multilingual variants of each input, collects independent judgments and rationales, and feeds inconsistent outputs back for self-reflection and re-evaluation. Experiments on multiple benchmarks show that SEMJ consistently outperforms voting and reflection baselines in both accuracy and cross-lingual consistency. Further analysis shows that inconsistency triggers useful re-evaluation, which improves judgment quality.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

ConSteer-RL: Steering Reasoning Capabilities in Large Language Models via Confidence-Aware Reinforcement Learning

Qing Miao, Yiming Zhao, Jing Yang et al. (+5 more)

Reinforcement Learning from Verifiable Rewards (RLVR) has recently become a key paradigm for improving the reasoning abilities of Large Language Models (LLMs), yet it remains limited by sparse binary rewards and its ignorance of model-internal uncertainty. In this paper, we propose ConSteer-RL, a simple yet effective framework that integrates token-level confidence signals derived from model log-probabilities into RLVR training. Specifically, building upon the Group Relative Policy Optimization (GRPO) framework, we construct a confidence-aware reward by aggregating per-token probabilities into a scalar confidence score and incorporating it into an awareness-based reward shaping mechanism that penalizes overconfident errors while reinforcing correct and confident reasoning. Experimental results demonstrate that ConSteer-RL consistently outperforms strong GRPO baselines, achieving average improvements of 2.3%-4.0% across different model scales.
Read the full abstract

All authors Qing Miao, Yiming Zhao, Jing Yang, Chenxi Liu, Yuehai Chen, Yuewen Liu, Shaoyi Du, Badong Chen

Computation and Language · Cross-listed from cs.SD

Assessing the Energy and Carbon Emissions of Neural Speaker Verification Model in Training and Inference

Hugo Leguillier, Driss Matrouf, Guillaume Lechien et al. (+1 more)

Deep-learning speaker verification (SV) increasingly relies on deep neural network backbones, whose environmental impact remains largely undocumented. In this paper, we conduct an evaluation of ResNet architectures trained on VoxCeleb2, varying depth, channel width, and stage distribution, and measure energy consumption and carbon footprint using node-level sensors. Results show a clear point of diminishing returns: deeper or wider models bring only marginal accuracy gains while energy consumption grows steeply. In contrast, mid-sized networks such as ResNet-50 and stage-concentrated variants achieve favorable trade-offs between performance and environmental impact. These findings provide actionable guidelines for designing energy-efficient SV systems.
Read the full abstract

All authors Hugo Leguillier, Driss Matrouf, Guillaume Lechien, Mickael Rouvier

Computation and Language

Aligned but Not Partner-Specific: How Multimodal LLM Agents Succeed in Reference Games Without Forming Conceptual Pacts

Po-Ya Angela Wang, Chinmaya Mishra, Asl{\i} Özyürek et al. (+2 more)

Repeated reference games test whether interlocutors replace their initially long descriptions with shorter, partner-specific expressions grounded in shared interaction history; that is, with conceptual pacts. Prior work shows that multimodal LLMs fail to become more efficient across rounds, although they align on the labels they use. However, how can we determine whether this alignment reflects partner-specific grounding rather than a shared task vocabulary? We address this by comparing competent multimodal agent dyads with human dyads from the KTH Tangrams corpus. Our novel methodological contribution is a pragmatically constrained pseudo-dyad baseline: rounds from two different real dyads describing the same target at comparable trajectory positions are paired, preserving referential task structure while removing shared partner history. This enables us to test whether the observed label alignment depends on interaction with a specific partner. Across three measures (task competence, description strategy, alignment dynamics), we find clear differences. Humans reduce effort through entrainment, compressing descriptions and increasing label alignment with partners. Agents instead maintain fixed effort levels, producing verbose descriptions from round one, with near-ceiling label overlap that is statistically indistinguishable between real and pseudo dyads. MLLMs thus achieve coordination without conceptual pacts, succeeding by verbose description rather than by forming the compact, history-dependent referring expressions characteristic of human dialogue.
Read the full abstract

All authors Po-Ya Angela Wang, Chinmaya Mishra, Asl{\i} Özyürek, Paula Rubio-Fernández, Esam Ghaleb

Computation and Language · Cross-listed from cs.SD

On Low-Bit Quantization Errors in Speaker Verification: Diagnostic and Mitigation

Hugo Leguillier, Driss Matrouf, Guillaume Lechien et al. (+1 more)

Although low-bit quantization provides practical means to deploy speaker verification on resource-constrained devices, its effects on speaker verification performance remain poorly understood. In this paper, we study uniform K-means quantization-aware training of ResNet-36 and ResNet-200 through joint layer-wise and score-level analyses. Our layer-wise analysis highlights fragile components and shows that score degradation is not fully explained by weight distortion alone. We identify a clear knee point at 2 bits, with larger score drift and harmful decision flips concentrated near the FP32 threshold. Our score-level analysis reveals where and how score errors emerge under extreme quantization. Building on these findings, we propose a calibrated multi-precision cascade that resolves most trials at 2 bits and escalates only ambiguous cases, achieving performance close to FP32 while preserving the efficiency benefits of low-bit inference with substantially lower compute and memory costs.
Read the full abstract

All authors Hugo Leguillier, Driss Matrouf, Guillaume Lechien, Mickael Rouvier

Computation and Language

Support Vector Rubrics: Closing the Gap Between Self-Generated and Human Rubrics

Mengyuan Sun, Yu Li, Zhuohao Yu et al. (+2 more)

Rubric-based evaluation is a promising paradigm for judging large language model (LLM) outputs, yet self-generated rubrics lag human-annotated criteria on hard instances. We argue this discriminative gap reflects an objective mismatch: self-generated rubrics describe good responses, whereas effective criteria must discriminate between close candidates. To close this gap, we introduce SVR (Support Vector Rubrics), a framework that recasts rubric construction as max-margin boundary learning over preference data. SVR mines contrastive features from preference pairs into a rubric bank, learns a prompt-conditioned selector together with global rubric weights, and iteratively refines the bank through support-pair selection and adversarial probing of hard negatives. At inference, given only the prompt, SVR retrieves the top-rubrics from the bank and scores responses. On RubricBench, SVR narrows the gap to human reference rubrics from 24.1 to 0.3 points and outperforms strong self-rubric and judge baselines, and the learned bank transfers across judges without retraining. On RewardBench 1&2, and RM-Bench, it remains competitive with dedicated reward models, demonstrating broader reward modeling capability. Overall, boundary-defining rubrics offer a principled route to closing the discriminative gap in LLM evaluation.
Read the full abstract

All authors Mengyuan Sun, Yu Li, Zhuohao Yu, Shikun Zhang, Wei Ye

Computation and Language

"I understand your perspective": LLM Persuasion through the Lens of Communicative Action Theory

Esra Dönmez and Agnieszka Falenska

Large Language Models (LLMs) can generate high-quality arguments, yet their ability to engage in nuanced and persuasive communicative actions remains largely unexplored. This work explores the persuasive potential of LLMs through the framework of Jürgen Habermas' Theory of Communicative Action. It examines whether LLMs express illocutionary intent (i.e., pragmatic functions of language such as conveying knowledge, building trust, or signaling similarity) in ways that are comparable to human communication. We simulate online discussions between opinion holders and LLMs using conversations from the persuasive subreddit ChangeMyView. We then compare the likelihood of illocutionary intents in human-written and LLM-generated counter-arguments, specifically those that successfully changed the original poster's view. We find that all three LLMs effectively convey illocutionary intent – often more so than humans – potentially increasing their anthropomorphism. Further, LLMs craft sycophantic responses that closely align with the opinion holder's intent, a strategy strongly associated with opinion change. Finally, crowd-sourced workers find LLM-generated counter-arguments more agreeable and consistently prefer them over human-written ones. These findings suggest that LLMs' persuasive power extends beyond merely generating high-quality arguments. On the contrary, training LLMs with human preferences effectively tunes them to mirror human communication patterns, particularly nuanced communicative actions, potentially increasing individuals' susceptibility to their influence.
Read the full abstract

Computation and Language

SurgiQ: A Large-Scale Multi-Domain Benchmark for Evaluating Surgical Understanding in Large Language Models

Ayah Al-Naji, Edoardo Fazzari, Saif Alkindi et al. (+3 more)

Reliable evaluation of large language models in surgery remains underdeveloped. Broad medical benchmarks test clinical knowledge, while surgery requires procedural reasoning, management trade-offs, negation handling, and selection among plausible operative decisions. We present SurgiQ, a text-only, source-grounded benchmark of 13,055 four-option multiple-choice questions spanning six surgical domains and four question formats: case-based, reasoning, best-option, and negative. SurgiQ is constructed from surgical textbooks, open-access papers, and examination material using a multi-stage generation, verification, and expert-audit pipeline. We evaluate 35 open-weight LLMs under a unified log-likelihood protocol. Our results show substantial remaining headroom: smaller models often remain near the 25% random baseline, while the best model reaches 68.1% accuracy. General-purpose models, especially Qwen2.5, outperform most biomedical models, suggesting that current medical specialization does not yet provide sufficiently broad surgical coverage. Calibration and error analysis further show that even strong models make confident mistakes on clinically plausible distractors, motivating more reliable and broader surgical LLM evaluation.
Read the full abstract

All authors Ayah Al-Naji, Edoardo Fazzari, Saif Alkindi, Hamdan Alhadhrami, Preslav Nakov, Cesare Stefanini

Computation and Language · Cross-listed from cs.CV

Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding?

Jiaqi Tang, Jianmin Chen, Youyang Zhai et al. (+6 more)

Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in visual understanding, yet their performance degrades significantly under real-world visual corruptions. While existing robustness enhancement approaches exist, they are limited: black-box feature alignment lacks interpretability, and white-box text-based reasoning cannot restore lost pixel-level details. This work investigates a fundamental research question: Can MLLMs recover corrupted visual content by themselves? To address this, we propose Robust-U1, a novel framework that equips MLLMs with explicit visual self-recovery capability for robust understanding. The approach comprises three core stages: supervised fine-tuning for initial reconstruction, reinforcement learning with dual rewards (pixel-level SSIM and semantic-level CLIP similarity) for aligning high visual quality, and multimodal reasoning that jointly considers both the corrupted input and the recovered image. Extensive experiments demonstrate that Robust-U1 achieves state-of-the-art robustness on the real-world corruption benchmark and maintains superior performance under adversarial corruptions on general VQA benchmarks. Analysis confirms that high-quality visual recovery directly enhances reasoning performance, establishing self-recovery as a critical mechanism for robust visual understanding. The source code is available at https://github.com/jqtangust/Robust-U1.
Read the full abstract

All authors Jiaqi Tang, Jianmin Chen, Youyang Zhai, Wei Wei, Runtao Liu, Mengjie Zhao, Xiangyu Wu, Qingfa Xiao, Qifeng Chen

Computation and Language

What's the Point? Spatial Grammar & Index Resolution for Sign Language Processing

Oline Ranum, Simon Hadfield and Richard Bowden

Sign language models are predominantly trained with gloss-sequence or text supervision, thereby under-modeling non-lexical and productive constructions. One comparatively tractable instance is spatial indexing: pointing gestures that assign discourse entities to spatial loci for subsequent co-reference, which lexicon-centric objectives largely fail to capture. We present a targeted evaluation of indexing in Sign Language Recognition, showing that despite comprising 10-15% of signing content, indexing is poorly recovered. We introduce a framework for training and evaluating indexing experts, establishing a baseline for index-aware sign language modeling. Our approach decomposes spatial reference resolution into index detection and discourse entity linking. The resulting mention representations enable automatic annotation and non-lexical structure modeling, and serve as an auxiliary indexing expert that augments a frozen SLR model at inference time.
Read the full abstract

Computation and Language

Diffusion Language Model Parallel Decoding via Product-of-Experts Bridge

Juntong Shi, Brian L. Trippe, Jure Leskovec et al. (+2 more)

Diffusion language models (DLMs) offer substantial speed advantages through parallel decoding, but the lack of token dependencies limits generation quality compared to autoregressive (AR) models. Recent progress attempts to bridge the gap via importance sampling, with DLM being the proposal and AR being the target. However, due to the huge gap between their distributions, the sampling requires a large number of particles and is thus expensive to compute. In this paper, we introduce PoE-Bridge, a novel decoding framework that drastically improves generation speed and accuracy by introducing an intermediate distribution to bridge the gap. The distribution is constructed as a Product-of-Experts (PoE) of the DLM proposal and the AR target. With the intermediate distribution, we first use the DLM to draft multiple continuations in parallel, then apply rejection sampling to verify the drafted tokens and move the resulting candidates toward the PoE. We then use importance sampling to further correct the PoE-aligned candidates toward the AR target. We further propose several improved techniques, including mixed-temperature sampling for enhanced diversity and elastic rejection windows for reducing wasted verification. Empirically, PoE-Bridge achieves significantly improved accuracy with $5\times$ speedup over the standard DLM decoding approach, and recovers at least 95% of the target AR model's performance, efficiently advancing most of the quality gap on challenging mathematical reasoning and coding tasks. Our code is available at https://github.com/juntongshi48/poe-bridge.
Read the full abstract

All authors Juntong Shi, Brian L. Trippe, Jure Leskovec, Stefano Ermon, Minkai Xu

Computation and Language · Cross-listed from cs.LG

When Behavioral Safety Evaluation Fails: A Representation-Level Perspective

Enyi Jiang, Anders Gjølbye, Yibo Jacky Zhang et al. (+1 more)

Safety evaluation of large language models (LLMs) is largely behavioral: a model is certified safe when it refuses harmful requests and answers benign ones. But refusing on the prompts an auditor happens to try does not show that the model is far from harmful behavior. Behavioral tests observe outputs; they do not measure how easily an intervention on the model turns a refusal into compliance. We call the gap between what static audits certify and what an intervention can reach the audit gap, and we show it is realizable: one can build a model that matches its safety-aligned base on every static audit yet gives way to a small, known perturbation of its internal state. We construct such dissociated models from three safety-aligned bases (Gemma 2 2B, Llama 3.2 3B, Qwen 2.5 3B) and audit the base, dissociated, and openly harmful models with the same soft interventions in parameter and latent space; the latent attacks are summarized by the Latent Vulnerability Score (LVS), the safety degradation produced per unit of bounded latent perturbation. Every static audit we run gives the dissociated model the same verdict as its base, since its refusals match the base, jailbreaks show no consistent signature, and a strong fixed probe on clean activations cannot tell it from the base. The same interventions an auditor could run reverse the verdict. At the targeted mid layer the dissociated models score 2.5 to 3.1 times higher LVS than their bases. A bounded latent attack elicits harmful compliance on 54 to 86% of prompts, against 3 to 48% for the bases, while matched random perturbations stay at or below 12%. Harmful fine-tuning reaches high compliance within five gradient steps, where the bases need 10 to 25. Behavioral testing, even with static latent probing, cannot certify representation-level robustness: a safety audit must intervene on the model, not only observe it.
Read the full abstract

All authors Enyi Jiang, Anders Gjølbye, Yibo Jacky Zhang, Sanmi Koyejo

Computation and Language · Cross-listed from cs.IR

GIScholarBench: Benchmarking LLM Overconfidence in GIS Research

Zongrng Li, Mingzheng Yang, Lei Zou et al. (+8 more)

Large language models (LLMs) are increasingly used in academic research workflows, but scholarly tasks require high factual precision and therefore expose a key weakness: overconfidence. Here, overconfidence is defined behaviorally as the tendency to produce confident, assertive, and well-formatted outputs even when the underlying knowledge is incomplete or unverifiable, rather than as a calibration gap between stated confidence and accuracy. To examine this issue, we introduce GIScholarBench, a benchmark built from 10,865 papers published in 25 core GIScience journals between 2020 and 2025. The benchmark covers three tasks with increasing cognitive complexity: metadata retrieval, literature linking, and research direction generation. We evaluate Claude Sonnet 4.5, Gemini 3, and ChatGPT 5.3 through their native web interfaces under real-world user-facing conditions. Results show consistent overconfidence across all tasks. In metadata retrieval, ChatGPT 5.3 achieves the highest accuracy, but all models still generate definitive titles and DOIs when predictions are wrong. In literature linking, Claude Sonnet 4.5 recovers the most references, but all models show a clear gap between top-ranked retrieval and longer citation lists, suggesting that references are extended beyond reliable retrieval capacity. In research direction generation, AI-generated directions show lower topic coverage, higher novel miss rates, and lower semantic diversity than real future-citing papers. These findings suggest that LLM overconfidence is task-invariant but takes different forms: factual overgeneration in retrieval, unreliable citation expansion in literature linking, and overconfidence in output completeness during research ideation.
Read the full abstract

All authors Zongrng Li, Mingzheng Yang, Lei Zou, Hongxu Ma, Hao Tian, Siqi Zhou, Wenjing Gong, Kaili Zhang, Bingqian Chen, Mitch Zhang, Yifan Yang

Computation and Language · Cross-listed from cs.CV

Sci-Rho: A Multilingual Visually-Grounded Symbolic Benchmark for STEM Problems

Muhammad Falensi Azmi, Ikhlasul Akmal Hanif, Vallerie Alexandra Putra et al. (+3 more)

Symbolic benchmarks have emerged as a key approach to assess model robustness under minor modifications to STEM-related questions. However, existing symbolic benchmarks mostly remain limited to mathematical reasoning, lack visual grounding, and are predominantly in English. In this work, we introduce Sci-Rho (Science Rhobustness), a dynamic benchmark for visually-grounded STEM problems spanning five subjects and seven languages, comprising 4,242 problem templates (606 per language) crafted by domain experts, including Olympiad medalists. Each template is implemented as executable Python code that generates diverse but equivalent problem instances by varying numerical values, visual patterns, geometric shapes, color schemes, and function types, resulting in 42,420 instances in total, each paired with reasoning steps and ground-truth solutions. We evaluated 17 state-of-the-art VLMs and discovered a noticeable gap between worst-case accuracy (defined as the proportion of problem templates that a model answers correctly across every generated variation) and average accuracy. We also discovered that smaller models show noticeable performance degradation across languages, whereas proprietary and larger models remain robust. Step-level evaluation reflects this same trend, revealing a significant gap between average F1 and worst-case F1 scores. Finally, our inspection of attention heads of a VLM reveals substantial cross-lingual variation in the relative attention allocated to image tokens compared to text tokens. Our work highlights the importance of evaluation beyond static benchmarks as a metric to measure the quality of VLMs.
Read the full abstract

All authors Muhammad Falensi Azmi, Ikhlasul Akmal Hanif, Vallerie Alexandra Putra, Adi Yeltay, Abdullah Mubarak, Fajri Koto

Computation and Language

Arabic Sentence Segmentation Across Genres and Punctuation Conditions

Mohammed Elkholy, Khalid N. Elmadani, Nizar Habash et al. (+1 more)

Sentence segmentation in Arabic is challenging due to ambiguous and inconsistent punctuation, with many texts lacking reliable sentence boundary markers. Existing approaches rely heavily on punctuation cues and are typically evaluated on well-formed text, limiting their robustness in realistic Arabic settings. To address this, we introduce AraSEG, a genre-diverse sentence segmentation corpus spanning eight genres and a wide range of punctuation and document structure conditions. Using AraSEG, we evaluate LLMs, lightweight encoder models, and dependency parser-based models under increasingly challenging segmentation settings. Our experiments show that lightweight encoders, and even dependency parser-based models, outperform LLMs under the hardest conditions. We further investigate the effects of training data size and genre diversity, finding that performance eventually saturates and cross-genre generalization remains challenging. We also demonstrate that accurate sentence segmentation substantially improves downstream dependency parsing. We make our code, data, and models publicly available.
Read the full abstract

All authors Mohammed Elkholy, Khalid N. Elmadani, Nizar Habash, Bashar Alhafni

Computation and Language · Cross-listed from cs.CV

IEA: Amateur-Friendly Conversational Image Editing Agent via Three Stages of Multitask Alignment

Zichen Zhu, Yuheng Sun, Mingxuan Zhu et al. (+11 more)

Current image editing software often hinges on fixed filters or expert tuning, leaving a gap between amateur users' intent and outcomes. Creations by generative models may contain artifacts, implausible details, or stylistic drift away from photorealism and offer little insight into why an edit was made. We propose IEA, a conversational Image Editing Agent that learns to operate parameterized tools in an explicit, interpretable action space. IEA is trained via a three-stage multitask pipeline: (1) SFT on distilled expert edits, (2) GRPO with rewards for likeness improvement, tool usefulness, and intent summarization, and (3) large-scale synthetic fine-tuning to jointly master image editing, refinement, and user intent summarization. By manipulating 16 editing tools step by step, IEA produces transparent edit traces that can be inspected and debugged. In quantitative experiments, it attains a lower pixel distance on the edit task and a higher ROUGE-L on the summary task than strong baselines. In user studies, it ranks best among tool-calling methods for instruction following while surpassing generative methods in overall perceptual quality. Our results validate interpretable, tool-centric VLMs as a reliable path to human instruction-guided image retouching.
Read the full abstract

All authors Zichen Zhu, Yuheng Sun, Mingxuan Zhu, Wenjie Ma, Situo Zhang, Zhexiang Wang, Ziyue Yang, Danyang Zhang, Kunyao Lan, Zihan Zhao, Dingye Liu, Siqi Xiang, Lu Chen, Kai Yu

Computation and Language

Rewrite to Translate, Translate to Reward: Reinforcement Learning for Source Rewriting in Machine Translation

Boxuan Lyu, Haiyue Song, Zhi Qu et al. (+3 more)

Prior work has explored prompting large language models (LLMs) to rewrite source text before translation, with the goal of improving machine translation (MT) quality. However, we find that such prompt-based rewriting can degrade translation quality rather than enhance it, particularly when smaller LLMs, such as 4B-parameter models, are used. We argue that this limitation stems from the difficulty of controlling rewriting behavior through natural-language prompts alone: a rewrite is useful only if it leads to a better downstream translation, yet existing prompt-based methods do not explicitly optimize for this signal. To address this issue, we propose RLSR (Reinforcement Learning for Source Rewriting), a reinforcement learning framework that trains the rewriting model with a reward derived from the downstream translation-quality improvement produced by each rewrite. Experiments across six MT models and 16 language pairs show that our 4B RLSR-trained rewriting models significantly outperform both the no-rewriting baseline and same-scale prompt-based rewriting baselines, while remaining competitive with baselines that use a 235B LLM. Our models and code are available at: https://github.com/vlaks425/MT-RLSR
Read the full abstract

All authors Boxuan Lyu, Haiyue Song, Zhi Qu, Hidetaka Kamigaito, Kotaro Funakoshi, Manabu Okumura

Computation and Language

Summarization is Not Dead Yet

Dongqi Liu, Chenxi Whitehouse, Zheng Zhao et al. (+3 more)

The progress of large language models (LLMs) has fueled claims that model-generated summaries rival or even surpass human-written references, raising questions about whether summarization remains an open research problem. We re-examine this narrative through a multi-track evaluation covering diverse datasets and state-of-the-art LLMs, combining controlled human assessment, bias-mitigated LLM-as-Judge protocols, factuality verification against external knowledge, and corpus-level linguistic analysis. Our findings reveal a more nuanced landscape in which human references continue to demonstrate advantages in informativeness and faithfulness, whereas LLM outputs are preferred mainly for surface-level coherence and fluency. Factuality verification indicates that human references remain more reliable, particularly for claims involving reasoning or synthesis, and linguistic analysis uncovers a pattern of stylistic homogeneity across different models. These observations suggest that current LLMs have raised the floor of summarization quality, but the ceiling of their performance remains below human capabilities.
Read the full abstract

All authors Dongqi Liu, Chenxi Whitehouse, Zheng Zhao, Zhuchen Cao, Jian Li, Yabiao Wang

Computation and Language

MC-PDD: Masked Corpus-Level Pretraining Data Detection for Black-Box Large Language Models

Kaixin Lan, Mu You, Tao Fang et al. (+3 more)

Pretraining is fundamental to the development of Large Language Models (LLMs), yet the opacity of pretraining data complicates model analysis and raises ethical, legal, and fairness concerns. Detecting whether specific datasets were used during pretraining is, therefore, critical. Existing state-of-the-art methods typically rely on access to model probability distributions, making them unsuitable for closed-source LLMs that provide only input-output interfaces. To address this limitation, we introduce Masked Corpus-level Pretraining Data Detection (MC-PDD), a novel method inspired by the masked language modeling paradigm. MC-PDD masks highly specific tokens in each text and prompts the LLM to predict the missing content. It then assesses whether the difference in prediction hit rates between a candidate corpus and a reference non-member corpus is statistically significant. Based on this comparison, MC-PDD determines whether the candidate texts were likely included in the model's pretraining data. Experimental results demonstrate clear and consistent differences in prediction hit rates between pretrained and unseen data across three datasets, for both open-source and closed-source LLMs. Despite operating under a stricter black-box setting, MC-PDD achieves performance comparable to existing detection methods. Our approach enables practical applications such as model auditing and data copyright verification using only standard API access. Upon acceptance, we will publicly release the code and datasets.
Read the full abstract

All authors Kaixin Lan, Mu You, Tao Fang, Binkai Ou, Lidia S. Chao, Derek F. Wong

Computation and Language

Customer-Agent: Overcoming Context Limitations in Ultra-Long Shopping Trajectories via Tool-Augmented Agents and RLVR

Hongye Liu, Rongmei Lin, Anurag Kashyap et al. (+4 more)

Understanding customer shopping trajectories is essential for enabling personalized shopping experiences. However, shopping records (i.e., customer's search, clicks, purchases, etc.) often span long time horizons over multiple years, resulting in extremely long trajectories that pose significant challenges for existing large language models (LLMs). Despite the importance of this problem, existing benchmarks are limited to short customer trajectories, while real-world trajectories from large e-commerce platforms are rarely accessible due to data privacy constraints. To address this gap, we introduce ShopTrajQA, a long-context evaluation benchmark constructed from real-world product information and simulated shopping trajectories. The dataset includes variants of up to 32k and 64k tokens, enabling systematic evaluation of model robustness under varying context lengths. Through comprehensive benchmarking of frontier LLMs, we identify critical performance gaps in reasoning over long shopping trajectory data. To address these challenges, we propose a Customer Agent Framework for ultra-long context management. Leveraging a Reinforcement Learning with Verifiable Rewards (RLVR) agentic training paradigm, our approach stores trajectories as external local files and trains the agent to autonomously retrieve and parse them through code-interpreter interactions (e.g., SQL queries), effectively bypassing the fixed in-context window constraints of LLMs. Experimental results demonstrate that our framework achieves strong performance for ShopTrajQA and shows generalization to other complex reasoning tasks.
Read the full abstract

All authors Hongye Liu, Rongmei Lin, Anurag Kashyap, Hejie Cui, Ricardo Henao, Besnik Fetahu, Bing Yin

Computation and Language · Cross-listed from cs.CV

FMRFusion: Frequency-Aware Multi-View Representation Learning for Heterogeneous Image Fusion

Tao Zhoua, Yunlong Liu, Qinghui Chen et al. (+6 more)

Infrared and visible image fusion aims to generate a composite image that retains significant target information and preserves detailed textures, integrating two heterogeneous modalities. Previous image fusion methods typically adopt a single-module stacking approach to extract features from the two modalities. However, these approaches may result in incomplete learning of their distinct characteristics, thereby limiting the fusion effectiveness and constrain ing robustness in real-world heterogeneous data scenarios. To address these challenges, we propose FMRFusion, a frequency-aware multi-view representation learning network for Heterogeneous Image Fusion. A Multi-Scale Struc tural Perception Module is introduced to effectively capture discriminative structures, extracting fine-grained local structures and essential contextual information. A bilinear frequency decomposition mechanism is employed to sepa rate features into high-frequency and low-frequency components, enabling joint modeling of local details and global representations across different frequency domains. Moreover, a Cross-View Complementary Interaction is incorpo rated to explicitly model and fuse the complementary characteristics between reflected light information and radiative intensity responses, facilitating effective cross-view interaction. We further improve the Performance of the fused results by flow matching, which progressively refines the fused features by learning the transformation from coarse data to high-quality representations. Extensive experiments conducted on multiple benchmark datasets demonstrate that FMRFusion achieves superior and consistent performance across a range of fusion tasks, especially in nighttime scenarios
Read the full abstract

All authors Tao Zhoua, Yunlong Liu, Qinghui Chen, Zekai Zhang, Minlong Sun, Changlin Biana, Dagang Li, Wenmin Wang, Jinglin Zhang

Computation and Language

MechLens: Late Crystallization of Factual Knowledge Explains Intervention Effectiveness in Language Models

Xueping Gao

Understanding where LLMs store factual knowledge is critical for hallucination mitigation. We systematically quantify Late Crystallization: factual knowledge does not gradually emerge across layers but "crystallizes" abruptly at the final layers. Across five model families (Pythia, Gemma, Qwen2.5, Llama-3.1, Mistral; 0.5–14B), 26.8%–93.4% of correct answers never enter top-10 predictions at any intermediate layer, with late emergence (>80% depth) consistent across architectures. Cross-scale (Qwen2.5-14B) and cross-benchmark (MMLU: 98.2%) results confirm generality; tuned lens rules out probe artifacts. A sentiment-classification control (0.5% for Qwen vs. 85.9% factual; 2.0% for Mistral vs. 26.8%) confirms the phenomenon is specific to factual recall. Late Crystallization yields a crystallization-guided intervention principle: CAA outperforms DoLa on moderate-crystallization models (Llama, Mistral; p<0.001), with a directionally consistent reversal on high-crystallization Qwen (+25.4% vs. +15.5% MC1, p=0.069). LayerNorm ablation shows crystallization is intrinsic to the residual stream; LN scaling (x1.2) yields +11.8% MC1 with zero inference overhead. We further reveal a Computability-Memorization Spectrum: computable knowledge crystallizes earlier (layer 22.1/28) than memorized facts (28.0/28). We release MechLens supporting five model families.
Read the full abstract

Computation and Language

Defending Against Malicious Finetuning by Scaling Train-time Adversarial Attacks

Haoming Wen, Shi Chen, Qingyu Shi et al. (+4 more)

Malicious finetuning attacks pose a major safety threat against open-weight large language models (LLMs). However, existing alignment-stage defenses provide limited protection against strong attacks that use full-parameter finetuning. To address this issue, we propose Patcher, a novel and efficient adversarial training algorithm that alternates between an attack stage and a defense stage: the attack stage simulates many-step malicious finetuning and computes the resulting parameter displacement as an "attack vector", while the defense stage aims to preserve safe behavior under such perturbations. Compared with other adversarial training methods, a key feature of Patcher is that the attack vector can be reused across multiple defense updates, substantially reducing training costs. Theoretically, we characterize how the relative gradient estimation error is influenced by the number of attack steps during training and reuse-duration of attack vectors. Empirically, experiment results show that Patcher reduces Attack Success Rate by 67.5%, 53.4% and 54.8% on three benchmarks compared to the second-best method while preserving the model's utility. Moreover, Patcher remains effective under diverse model sizes and attack scenarios, and generalizes to prompt-based jailbreak attacks. Code is available at https://github.com/haomingwen/patcher
Read the full abstract

All authors Haoming Wen, Shi Chen, Qingyu Shi, Siyuan Liu, Minrui Luo, Jingzhao Zhang, Tianxing He

Computation and Language

Neutrality Bites: Gender Representation in AI-Generated Animal Stories

Imani Finkley, Yuanxi Li and Melanie Walsh

Gender bias in AI-generated stories is a well-documented problem. While much attention has been paid to reducing or mitigating this bias, it is not always clear whether interventions produce genuinely fairer results. To investigate this issue, we examine how large language models (LLMs) handle gender assignment in a narrative context that is popular, highly ambiguous, and also known to closely reproduce human stereotypes: stories about talking animals. We prompt six leading LLMs to complete an English-language story about seven different anthropomorphic animal characters whose gender is unstated. We additionally iterate with four different narrative settings and a range of model temperatures. Across the 23.8K stories, we find that models frequently avoid gendering the animal character in the story (19% on average) or use gender-neutral language like "it" or "its" (38.2% on average). However, when gender is assigned, there is a significant masculine bias. Feminine animal characters are virtually absent, present in just 2.2% of stories vs. 40.6% that feature masculine characters. Our findings point to a broader argument: neutrality bites. In other words, models that prioritize neutrality to address social bias may actually contribute to the erasure of marginalized perspectives and identities. We suggest that alternative strategies beyond neutrality need to be pursued, such as ones that more equally distribute social possibilities across imagined subjects.
Read the full abstract

Computation and Language

What Does Debiasing Really Remove? A Geometric Study of PCA-Based Gender Debiasing in Word Embeddings

Alexey Kresin, Tchifou M. Dieffi and Tomer Caspi

Debiasing methods based on principal component analysis (PCA) are broadly used to reduce gender bias in word embeddings used in LLMs, yet it remains unclear what aspects of bias they actually remove and how destructive this process is. These methods are based on the understanding that bias resides in a low-dimensional subspace, with the assumption that most of it can be captured by a few principal components. In this work, we conduct a systematic geometric analysis of PCA-based gender debiasing and investigate what is actually removed from the embedding space. Our experiments across multiple embeddings show that direct gender bias is primarily concentrated in the first principal component, supporting the low-rank bias hypothesis. However, associative bias measured by WEAT does not align with these principal directions and is instead spread across multiple embedding dimensions. Furthermore, as expected, we demonstrate that removing an increasing number of principal components leads to a consistent degradation of the embedding geometry, affecting semantic structure and vector relationships. These results reveal that PCA-based debiasing operates as a trade-off: while it effectively reduces certain forms of direct bias, it fails to eliminate distributed associations and introduces geometric distortion. Moreover, there is no universal optimal level of debiasing, as the balance between bias reduction and semantic preservation depends on the chosen metric and embedding. Overall, our findings suggest that bias in word embeddings is not purely low-rank and that simple subspace removal methods may be insufficient for comprehensive debiasing.
Read the full abstract