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

Thursday 8 January

Computation and Language

On the Limitations of Rank-One Model Editing in Answering Multi-hop Questions

Zhiyuan He, Binghan Chen, Tianxiang Xiong et al. (+3 more)

Recent advances in Knowledge Editing (KE), particularly Rank-One Model Editing (ROME), show superior efficiency over fine-tuning and in-context learning for updating single-hop facts in transformers. However, these methods face significant challenges when applied to multi-hop reasoning tasks requiring knowledge chaining. In this work, we study the effect of editing knowledge with ROME on different layer depths and identify three key failure modes. First, the "hopping-too-late" problem occurs as later layers lack access to necessary intermediate representations. Second, generalization ability deteriorates sharply when editing later layers. Third, the model overfits to edited knowledge, incorrectly prioritizing edited-hop answers regardless of context. To mitigate the issues of "hopping-too-late" and generalisation decay, we propose Redundant Editing, a simple yet effective strategy that enhances multi-hop reasoning. Our experiments demonstrate that this approach can improve accuracy on 2-hop questions by at least 15.5 percentage points, representing a 96% increase over the previous single-edit strategy, while trading off some specificity and language naturalness.
Read the full abstract

All authors Zhiyuan He, Binghan Chen, Tianxiang Xiong, Ziyang Sun, Mozhao Zhu, Xi Chen

Computation and Language

THaLLE-ThaiLLM: Domain-Specialized Small LLMs for Finance and Thai – Technical Report

KBTG Labs: Anuruth Lertpiya, Danupat Khamnuansin, Kantapong Sucharitpongpan et al. (+4 more)

Large Language Models (LLMs) have demonstrated significant potential across various domains, particularly in banking and finance, where they can automate complex tasks and enhance decision-making at scale. Due to privacy, security, and regulatory concerns, organizations often prefer on-premise deployment of LLMs. The ThaiLLM initiative aims to enhance Thai language capabilities in open-LLMs, enabling Thai industry to leverage advanced language models. However, organizations often face a trade-off between deploying multiple specialized models versus the prohibitive expense of training a single multi-capability model. To address this, we explore model merging as a resource-efficient alternative for developing high-performance, multi-capability LLMs. We present results from two key experiments: first, merging Qwen-8B with ThaiLLM-8B demonstrates how ThaiLLM-8B enhances Thai general capabilities, showing an uplift of M3 and M6 O-NET exams over the general instruction-following Qwen-8B. Second, we merge Qwen-8B with both ThaiLLM-8B and THaLLE-CFA-8B. This combination results in further improvements in performance across both general and financial domains, by demonstrating an uplift in both M3 and M6 O-NET, Flare-CFA, and Thai-IC benchmarks. The report showcases the viability of model merging for efficiently creating multi-capability LLMs.
Read the full abstract

All authors KBTG Labs: Anuruth Lertpiya, Danupat Khamnuansin, Kantapong Sucharitpongpan, Pornchanan Balee, Tawunrat Chalothorn, Thadpong Pongthawornkamol, Monchai Lertsutthiwong

Computation and Language

Aligning Text, Code, and Vision: A Multi-Objective Reinforcement Learning Framework for Text-to-Visualization

Mizanur Rahman, Mohammed Saidul Islam, Md Tahmid Rahman Laskar et al. (+2 more)

Text-to-Visualization (Text2Vis) systems translate natural language queries over tabular data into concise answers and executable visualizations. While closed-source LLMs generate functional code, the resulting charts often lack semantic alignment and clarity, qualities that can only be assessed post-execution. Open-source models struggle even more, frequently producing non-executable or visually poor outputs. Although supervised fine-tuning can improve code executability, it fails to enhance overall visualization quality, as traditional SFT loss cannot capture post-execution feedback. To address this gap, we propose RL-Text2Vis, the first reinforcement learning framework for Text2Vis generation. Built on Group Relative Policy Optimization (GRPO), our method uses a novel multi-objective reward that jointly optimizes textual accuracy, code validity, and visualization quality using post-execution feedback. By training Qwen2.5 models (7B and 14B), RL-Text2Vis achieves a 22% relative improvement in chart quality over GPT-4o on the Text2Vis benchmark and boosts code execution success from 78% to 97% relative to its zero-shot baseline. Our models significantly outperform strong zero-shot and supervised baselines and also demonstrate robust generalization to out-of-domain datasets like VIS-Eval and NVBench. These results establish GRPO as an effective strategy for structured, multimodal reasoning in visualization generation. We release our code at https://github.com/vis-nlp/RL-Text2Vis.
Read the full abstract

All authors Mizanur Rahman, Mohammed Saidul Islam, Md Tahmid Rahman Laskar, Shafiq Joty, Enamul Hoque

Computation and Language

FeedEval: Pedagogically Aligned Evaluation of LLM-Generated Essay Feedback

Seongyeub Chu, Jongwoo Kim and Munyong Yi

Going beyond the prediction of numerical scores, recent research in automated essay scoring has increasingly emphasized the generation of high-quality feedback that provides justification and actionable guidance. To mitigate the high cost of expert annotation, prior work has commonly relied on LLM-generated feedback to train essay assessment models. However, such feedback is often incorporated without explicit quality validation, resulting in the propagation of noise in downstream applications. To address this limitation, we propose FeedEval, an LLM-based framework for evaluating LLM-generated essay feedback along three pedagogically grounded dimensions: specificity, helpfulness, and validity. FeedEval employs dimension-specialized LLM evaluators trained on datasets curated in this study to assess multiple feedback candidates and select high-quality feedback for downstream use. Experiments on the ASAP++ benchmark show that FeedEval closely aligns with human expert judgments and that essay scoring models trained with FeedEval-filtered high-quality feedback achieve superior scoring performance. Furthermore, revision experiments using small LLMs show that the high-quality feedback identified by FeedEval leads to more effective essay revisions. We release our code and curated datasets at: https://github.com/BBeeChu/FeedEval.git.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Neurosymbolic Retrievers for Retrieval-augmented Generation

Yash Saxena and Manas Gaur

Retrieval Augmented Generation (RAG) has made significant strides in overcoming key limitations of large language models, such as hallucination, lack of contextual grounding, and issues with transparency. However, traditional RAG systems consist of three interconnected neural components - the retriever, re-ranker, and generator - whose internal reasoning processes remain opaque. This lack of transparency complicates interpretability, hinders debugging efforts, and erodes trust, especially in high-stakes domains where clear decision-making is essential. To address these challenges, we introduce the concept of Neurosymbolic RAG, which integrates symbolic reasoning using a knowledge graph with neural retrieval techniques. This new framework aims to answer two primary questions: (a) Can retrievers provide a clear and interpretable basis for document selection? (b) Can symbolic knowledge enhance the clarity of the retrieval process? We propose three methods to improve this integration. First is MAR (Knowledge Modulation Aligned Retrieval) that employs modulation networks to refine query embeddings using interpretable symbolic features, thereby making document matching more explicit. Second, KG-Path RAG enhances queries by traversing knowledge graphs to improve overall retrieval quality and interpretability. Lastly, Process Knowledge-infused RAG utilizes domain-specific tools to reorder retrieved content based on validated workflows. Preliminary results from mental health risk assessment tasks indicate that this neurosymbolic approach enhances both transparency and overall performance
Read the full abstract

Computation and Language · Cross-listed from cs.AI

BackdoorAgent: A Unified Framework for Backdoor Attacks on LLM-based Agents

Yunhao Feng, Yige Li, Yutao Wu et al. (+6 more)

Large language model (LLM) agents execute tasks through multi-step workflows that combine planning, memory, and tool use. While this design enables autonomy, it also expands the attack surface for backdoor threats. Backdoor triggers injected into specific stages of an agent workflow can persist through multiple intermediate states and adversely influence downstream outputs. However, existing studies remain fragmented and typically analyze individual attack vectors in isolation, leaving the cross-stage interaction and propagation of backdoor triggers poorly understood from an agent-centric perspective. To fill this gap, we propose BackdoorAgent, a modular and stage-aware framework that provides a unified, agent-centric view of backdoor threats in LLM agents. BackdoorAgent structures the attack surface into three functional stages of agentic workflows, including planning attacks, memory attacks, and tool-use attacks, and instruments agent execution to enable systematic analysis of trigger activation and propagation across different stages. Building on this framework, we construct a standardized benchmark spanning four representative agent applications: Agent QA, Agent Code, Agent Web, and Agent Drive, covering both language-only and multimodal settings. Our empirical analysis shows that triggers implanted at a single stage can persist across multiple steps and propagate through intermediate states. For instance, when using a GPT-based backbone, we observe trigger persistence in 43.58% of planning attacks, 77.97% of memory attacks, and 60.28% of tool-stage attacks, highlighting the vulnerabilities of the agentic workflow itself to backdoor threats. To facilitate reproducibility and future research, our code and benchmark are publicly available at GitHub.
Read the full abstract

All authors Yunhao Feng, Yige Li, Yutao Wu, Yingshui Tan, Yanming Guo, Yifan Ding, Kun Zhai, Xingjun Ma, Yu-Gang Jiang

Computation and Language · Cross-listed from cs.LG

A Vision for Multisensory Intelligence: Sensing, Science, and Synergy

Paul Pu Liang

Our experience of the world is multisensory, spanning a synthesis of language, sight, sound, touch, taste, and smell. Yet, artificial intelligence has primarily advanced in digital modalities like text, vision, and audio. This paper outlines a research vision for multisensory artificial intelligence over the next decade. This new set of technologies can change how humans and AI experience and interact with one another, by connecting AI to the human senses and a rich spectrum of signals from physiological and tactile cues on the body, to physical and social signals in homes, cities, and the environment. We outline how this field must advance through three interrelated themes of sensing, science, and synergy. Firstly, research in sensing should extend how AI captures the world in richer ways beyond the digital medium. Secondly, developing a principled science for quantifying multimodal heterogeneity and interactions, developing unified modeling architectures and representations, and understanding cross-modal transfer. Finally, we present new technical challenges to learn synergy between modalities and between humans and AI, covering multisensory integration, alignment, reasoning, generation, generalization, and experience. Accompanying this vision paper are a series of projects, resources, and demos of latest advances from the Multisensory Intelligence group at the MIT Media Lab, see https://mit-mi.github.io/.
Read the full abstract

Computation and Language

Identifying Good and Bad Neurons for Task-Level Controllable LLMs

Wenjie Li, Guansong Pang, Hezhe Qiao et al. (+2 more)

Large Language Models have demonstrated remarkable capabilities on multiple-choice question answering benchmarks, but the complex mechanisms underlying their large-scale neurons remain opaque, posing significant challenges for understanding and steering LLMs. While recent studies made progress on identifying responsible neurons for certain abilities, these ability-specific methods are infeasible for task-focused scenarios requiring coordinated use of multiple abilities. Moreover, these approaches focus only on supportive neurons that correlate positively with task completion, while neglecting neurons with other roles-such as inhibitive roles-and misled neuron attribution due to fortuitous behaviors in LLMs (i.e., correctly answer the questions by chance rather than genuine understanding). To address these challenges, we propose NeuronLLM, a novel task-level LLM understanding framework that adopts the biological principle of functional antagonism for LLM neuron identification. The key insight is that task performance is jointly determined by neurons with two opposing roles: good neurons that facilitate task completion and bad neurons that inhibit it. NeuronLLM achieves a holistic modeling of neurons via contrastive learning of good and bad neurons, while leveraging augmented question sets to mitigate the fortuitous behaviors in LLMs. Comprehensive experiments on LLMs of different sizes and families show the superiority of NeuronLLM over existing methods in four NLP tasks, providing new insights into LLM functional organization.
Read the full abstract

All authors Wenjie Li, Guansong Pang, Hezhe Qiao, Debin Gao, David Lo

Computation and Language · Cross-listed from cs.LG

Linear Dynamics in the RLVR Training of Large Language Models

Tianle Wang, Jiayu Liu, Zhongyuan Wu et al. (+4 more)

Reinforcement learning with verifiable rewards (RLVR) has driven significant performance gains in reasoning-oriented large language models (LLMs), yet its internal training dynamics remain largely a black box. In this work, we perform a comprehensive trajectory-level analysis of RLVR and uncover a striking regularity: across various model families, RL algorithms, and training configurations, RLVR consistently enters a robust linear regime, where both parameter weights and output log-probabilities, measured rigorously via teacher-forced evaluation, evolve in a highly linear manner ($R^2 > 0.7$). Through controlled experiments and theoretical analysis, we demonstrate that this linearity is not a coincidence, but stems from the high-variance, noisy nature of RLVR training signals, which act as a low-pass filter to concentrate optimization along a stable, low-dimensional drift. Moreover, we show that this linear structure is not merely descriptive but powerfully predictive and actionable. Specifically, weight-space extrapolation matches the performance of standard RL optimization while achieving a 6.1x training speedup through periodic re-grounding. Meanwhile, output-space extrapolation serves as a lightweight intervention that effectively bypasses late-stage model collapse, consistently outperforming standard RL across mathematical and coding benchmarks, with an average performance improvement of 4.2%. Our code is available at https://github.com/Miaow-Lab/RLVR-Linearity.
Read the full abstract

All authors Tianle Wang, Jiayu Liu, Zhongyuan Wu, Shenghao Jin, Wei Chen, Hao Xu, Ning Miao

Computation and Language

BanglaLorica: Design and Evaluation of a Robust Watermarking Algorithm for Large Language Models in Bangla Text Generation

Amit Bin Tariqul, A N M Zahid Hossain Milkan, Sahab-Al-Chowdhury et al. (+3 more)

As large language models (LLMs) are increasingly deployed for text generation, watermarking has become essential for authorship attribution, intellectual property protection, and misuse detection. While existing watermarking methods perform well in high-resource languages, their robustness in low-resource languages remains underexplored. This work presents the first systematic evaluation of state-of-the-art text watermarking methods: KGW, Exponential Sampling (EXP), and Waterfall, for Bangla LLM text generation under cross-lingual round-trip translation (RTT) attacks. Under benign conditions, KGW and EXP achieve high detection accuracy (>88%) with negligible perplexity and ROUGE degradation. However, RTT causes detection accuracy to collapse below RTT causes detection accuracy to collapse to 9-13%, indicating a fundamental failure of token-level watermarking. To address this, we propose a layered watermarking strategy that combines embedding-time and post-generation watermarks. Experimental results show that layered watermarking improves post-RTT detection accuracy by 25-35%, achieving 40-50% accuracy, representing a 3$\times$ to 4$\times$ relative improvement over single-layer methods, at the cost of controlled semantic degradation. Our findings quantify the robustness-quality trade-off in multilingual watermarking and establish layered watermarking as a practical, training-free solution for low-resource languages such as Bangla. Our code and data will be made public.
Read the full abstract

All authors Amit Bin Tariqul, A N M Zahid Hossain Milkan, Sahab-Al-Chowdhury, Syed Rifat Raiyan, Hasan Mahmud, Md Kamrul Hasan

Computation and Language · Cross-listed from cs.SE

Advancing Language Models for Code-related Tasks

Zhao Tian

Recent advances in language models (LMs) have driven significant progress in various software engineering tasks. However, existing LMs still struggle with complex programming scenarios due to limitations in data quality, model architecture, and reasoning capability. This research systematically addresses these challenges through three complementary directions: (1) improving code data quality with a code difference-guided adversarial augmentation technique (CODA) and a code denoising technique (CodeDenoise); (2) enhancing model architecture via syntax-guided code LMs (LEAM and LEAM++); and (3) advancing model reasoning with a prompting technique (muFiX) and an agent-based technique (Specine). These techniques aim to promote the practical adoption of LMs in software development and further advance intelligent software engineering.
Read the full abstract

Computation and Language

GRACE: Reinforcement Learning for Grounded Response and Abstention under Contextual Evidence

Yibo Zhao, Jiapeng Zhu, Zichen Ding et al. (+1 more)

Retrieval-Augmented Generation (RAG) integrates external knowledge to enhance Large Language Models (LLMs), yet systems remain susceptible to two critical flaws: providing correct answers without explicit grounded evidence and producing fabricated responses when the retrieved context is insufficient. While prior research has addressed these issues independently, a unified framework that integrates evidence-based grounding and reliable abstention is currently lacking. In this paper, we propose GRACE, a reinforcement-learning framework that simultaneously mitigates both types of flaws. GRACE employs a data construction method that utilizes heterogeneous retrievers to generate diverse training samples without manual annotation. A multi-stage gated reward function is then employed to train the model to assess evidence sufficiency, extract key supporting evidence, and provide answers or explicitly abstain. Experimental results on two benchmarks demonstrate that GRACE achieves state-of-the-art overall accuracy and strikes a favorable balance between accurate response and rejection, while requiring only 10% of the annotation costs of prior methods. Our code is available at https://github.com/YiboZhao624/Grace..
Read the full abstract

All authors Yibo Zhao, Jiapeng Zhu, Zichen Ding, Xiang Li

Computation and Language

LinguaGame: A Linguistically Grounded Game-Theoretic Paradigm for Multi-Agent Dialogue Generation

Yuxiao Ye, Yiming Zhang, Yiran Ma et al. (+3 more)

Large Language Models (LLMs) have enabled Multi-Agent Systems (MASs) where agents interact through natural language to solve complex tasks or simulate multi-party dialogues. Recent work on LLM-based MASs has mainly focused on architecture design, such as role assignment and workflow orchestration. In contrast, this paper targets the interaction process itself, aiming to improve agents' communication efficiency by helping them convey their intended meaning more effectively through language. To this end, we propose LinguaGame, a linguistically-grounded game-theoretic paradigm for multi-agent dialogue generation. Our approach models dialogue as a signalling game over communicative intents and strategies, solved with a training-free equilibrium approximation algorithm for inference-time decision adjustment. Unlike prior game-theoretic MASs, whose game designs are often tightly coupled with task-specific objectives, our framework relies on linguistically informed reasoning with minimal task-specific coupling. Specifically, it treats dialogue as intentional and strategic communication, requiring agents to infer what others aim to achieve (intents) and how they pursue those goals (strategies). We evaluate our framework in simulated courtroom proceedings and debates, with human expert assessments showing significant gains in communication efficiency.
Read the full abstract

All authors Yuxiao Ye, Yiming Zhang, Yiran Ma, Huiyuan Xie, Huining Zhu, Zhiyuan Liu

Computation and Language

WESR: Scaling and Evaluating Word-level Event-Speech Recognition

Chenchen Yang, Kexin Huang, Liwei Fan et al. (+8 more)

Speech conveys not only linguistic information but also rich non-verbal vocal events such as laughing and crying. While semantic transcription is well-studied, the precise localization of non-verbal events remains a critical yet under-explored challenge. Current methods suffer from insufficient task definitions with limited category coverage and ambiguous temporal granularity. They also lack standardized evaluation frameworks, hindering the development of downstream applications. To bridge this gap, we first develop a refined taxonomy of 21 vocal events, with a new categorization into discrete (standalone) versus continuous (mixed with speech) types. Based on the refined taxonomy, we introduce WESR-Bench, an expert-annotated evaluation set (900+ utterances) with a novel position-aware protocol that disentangles ASR errors from event detection, enabling precise localization measurement for both discrete and continuous events. We also build a strong baseline by constructing a 1,700+ hour corpus, and train specialized models, surpassing both open-source audio-language models and commercial APIs while preserving ASR quality. We anticipate that WESR will serve as a foundational resource for future research in modeling rich, real-world auditory scenes.
Read the full abstract

All authors Chenchen Yang, Kexin Huang, Liwei Fan, Qian Tu, Botian Jiang, Dong Zhang, Linqi Yin, Shimin Li, Zhaoye Fei, Qinyuan Cheng, Xipeng Qiu

Computation and Language · Cross-listed from cs.AI

CircuitLM: A Multi-Agent LLM-Aided Design Framework for Generating Circuit Schematics from Natural Language Prompts

Khandakar Shakib Al Hasan, Syed Rifat Raiyan, Hasin Mahtab Alvee et al. (+1 more)

Generating accurate circuit schematics from high-level natural language descriptions remains a persistent challenge in electronic design automation (EDA), as large language models (LLMs) frequently hallucinate components, violate strict physical constraints, and produce non-machine-readable outputs. To address this, we present CircuitLM, a multi-agent pipeline that translates user prompts into structured, visually interpretable $\texttt{CircuitJSON}$ schematics. The framework mitigates hallucination and ensures physical viability by grounding generation in a curated, embedding-powered component knowledge base through five sequential stages: (i) component identification, (ii) canonical pinout retrieval, (iii) chain-of-thought reasoning, (iv) JSON schematic synthesis, and (v) interactive force-directed visualization. We evaluate the system on a dataset of 100 unique circuit-design prompts using five state-of-the-art LLMs. To systematically assess performance, we deploy a rigorous dual-layered evaluation methodology: a deterministic Electrical Rule Checking (ERC) engine categorizes topological faults by strict severity (Critical, Major, Minor, Warning), while an LLM-as-a-judge meta-evaluator identifies complex, context-aware design flaws that bypass standard rule-based checkers. Ultimately, this work demonstrates how targeted retrieval combined with deterministic and semantic verification can bridge natural language to structurally viable, schematic-ready hardware and safe circuit prototyping. Our code and data are publicly available at https://github.com/Khandakar227/CircuitLM.
Read the full abstract

All authors Khandakar Shakib Al Hasan, Syed Rifat Raiyan, Hasin Mahtab Alvee, Wahid Sadik

Computation and Language · Cross-listed from cs.CV

Vision-Language Agents for Interactive Forest Change Analysis

James Brock, Ce Zhang and Nantheera Anantrasirichai

Modern forest monitoring workflows increasingly benefit from the growing availability of high-resolution satellite imagery and advances in deep learning. Two persistent challenges in this context are accurate pixel-level change detection and meaningful semantic change captioning for complex forest dynamics. While large language models (LLMs) are being adapted for interactive data exploration, their integration with vision-language models (VLMs) for remote sensing image change interpretation (RSICI) remains underexplored. To address this gap, we introduce an LLM-driven agent for integrated forest change analysis that supports natural language querying across multiple RSICI tasks. The proposed system builds upon a multi-level change interpretation (MCI) vision-language backbone with LLM-based orchestration. To facilitate adaptation and evaluation in forest environments, we further introduce the Forest-Change dataset, which comprises bi-temporal satellite imagery, pixel-level change masks, and multi-granularity semantic change captions generated using a combination of human annotation and rule-based methods. Experimental results show that the proposed system achieves mIoU and BLEU-4 scores of 67.10% and 40.17% on the Forest-Change dataset, and 88.13% and 34.41% on LEVIR-MCI-Trees, a tree-focused subset of LEVIR-MCI benchmark for joint change detection and captioning. These results highlight the potential of interactive, LLM-driven RSICI systems to improve accessibility, interpretability, and efficiency of forest change analysis. All data and code are publicly available at https://github.com/JamesBrockUoB/ForestChat.
Read the full abstract

Computation and Language

SampoNLP: A Self-Referential Toolkit for Morphological Analysis of Subword Tokenizers

Iaroslav Chelombitko, Ekaterina Chelombitko and Aleksey Komissarov

The quality of subword tokenization is critical for Large Language Models, yet evaluating tokenizers for morphologically rich Uralic languages is hampered by the lack of clean morpheme lexicons. We introduce SampoNLP, a corpus-free toolkit for morphological lexicon creation using MDL-inspired Self-Referential Atomicity Scoring, which filters composite forms through internal structural cues - suited for low-resource settings. Using the high-purity lexicons generated by SampoNLP for Finnish, Hungarian, and Estonian, we conduct a systematic evaluation of BPE tokenizers across a range of vocabulary sizes (8k-256k). We propose a unified metric, the Integrated Performance Score (IPS), to navigate the trade-off between morpheme coverage and over-splitting. By analyzing the IPS curves, we identify the "elbow points" of diminishing returns and provide the first empirically grounded recommendations for optimal vocabulary sizes (k) in these languages. Our study not only offers practical guidance but also quantitatively demonstrates the limitations of standard BPE for highly agglutinative languages. The SampoNLP library and all generated resources are made publicly available: https://github.com/AragonerUA/SampoNLP
Read the full abstract

Computation and Language

Concept Tokens: Learning Behavioral Embeddings Through Concept Definitions

Ignacio Sastre and Aiala Rosá

We propose Concept Tokens, a lightweight method that adds a new special token to a pretrained LLM and learns only its embedding from multiple natural language definitions of a target concept, where occurrences of the concept are replaced by the new token. The LLM is kept frozen and the embedding is optimized with the standard language-modeling objective. We evaluate Concept Tokens in three settings. First, we study hallucinations in closed-book question answering on HotpotQA and find a directional effect: negating the hallucination token reduces hallucinated answers mainly by increasing abstentions, whereas asserting it increases hallucinations and lowers precision. Second, we induce recasting, a pedagogical feedback strategy for second language teaching, and observe the same directional effect. Moreover, compared to providing the full definitional corpus in-context, concept tokens better preserve compliance with other instructions (e.g., asking follow-up questions). Finally, we include a qualitative study with the Eiffel Tower and a fictional "Austral Tower" to illustrate what information the learned embeddings capture and where their limitations emerge. Overall, Concept Tokens provide a compact control signal learned from definitions that can steer behavior in frozen LLMs.
Read the full abstract

Computation and Language

Beyond Static Summarization: Proactive Memory Extraction for LLM Agents

Chengyuan Yang, Zequn Sun, Wei Wei et al. (+1 more)

Memory management is vital for LLM agents in long-term and personalized interactions. Most previous work studies how to retrieve and use memory, but pays less attention to how memory is extracted. We find two main limitations in existing methods. First, extraction is "ahead-of-time": the agent saves information before it knows future tasks. A single summary prompt often mixes details, events, and relations, so useful information is lost. Second, extraction is usually one-off. Without verification, errors and hallucinations may stay in memory for a long time. To address these limitations, we propose ProMem, a proactive memory extraction framework. It separates details, events, and relations, and uses different extraction strategies for each type. It also checks completeness to recover missed events and verifies facts at the atomic level to reduce hallucinations. Experiments show that ProMem improves memory completeness and QA accuracy, while keeping a good balance between quality and token cost.
Read the full abstract

All authors Chengyuan Yang, Zequn Sun, Wei Wei, Wei Hu

Computation and Language

Users Mispredict Their Own Preferences for AI Writing Assistance

Vivian Lai, Zana Buçinca, Nil-Jana Akpinar et al. (+5 more)

Proactive AI writing assistants need to predict when users want drafting help, yet we lack empirical understanding of what drives preferences. Through a factorial vignette study with 50 participants making 750 pairwise comparisons, we find compositional effort dominates decisions ($\rho = 0.597$) while urgency shows no predictive power ($\rho \approx 0$). More critically, users exhibit a striking perception-behavior gap: they rank urgency first in self-reports despite it being the weakest behavioral driver, representing a complete preference inversion. This misalignment has measurable consequences. Systems designed from users' stated preferences achieve only 57.7% accuracy, underperforming even naive baselines, while systems using behavioral patterns reach significantly higher 61.3% ($p < 0.05$). These findings demonstrate that relying on user introspection for system design actively misleads optimization, with direct implications for proactive natural language generation (NLG) systems.
Read the full abstract

All authors Vivian Lai, Zana Buçinca, Nil-Jana Akpinar, Mo Houtti, Hyeonsu B. Kang, Kevin Chian, Namjoon Suh, Alex C. Williams

Computation and Language · Cross-listed from cs.IR

Re-Rankers as Relevance Judges

Chuan Meng, Jiqun Liu, Mohammad Aliannejadi et al. (+3 more)

Using large language models (LLMs) to predict relevance judgments has shown promising results. Most studies treat this task as a distinct research line, e.g., focusing on prompt design for predicting relevance labels given a query and passage. However, predicting relevance judgments is essentially a form of relevance prediction, a problem extensively studied in tasks such as re-ranking. Despite this potential overlap, little research has explored reusing or adapting established re-ranking methods to predict relevance judgments, leading to potential resource waste and redundant development. To bridge this gap, we reproduce re-rankers in a re-ranker-as-relevance-judge setup. We design two adaptation strategies: (i) using binary tokens (e.g., "true" and "false") generated by a re-ranker as direct judgments, and (ii) converting continuous re-ranking scores into binary labels via thresholding. We perform extensive experiments on TREC-DL 2019 to 2023 with 8 re-rankers from 3 families, ranging from 220M to 32B, and analyse the evaluation bias exhibited by re-ranker-based judges. Results show that re-ranker-based relevance judges, under both strategies, can outperform UMBRELA, a state-of-the-art LLM-based relevance judge, in around 40% to 50% of the cases; they also exhibit strong self-preference towards their own and same-family re-rankers, as well as cross-family bias.
Read the full abstract

All authors Chuan Meng, Jiqun Liu, Mohammad Aliannejadi, Fengran Mo, Jeff Dalton, Maarten de Rijke

Wednesday 7 January

Computation and Language

Retrieval Heads are Dynamic

Yuping Lin, Zitao Li, Yue Xing et al. (+6 more)

Recent studies have identified "retrieval heads" in Large Language Models (LLMs) responsible for extracting information from input contexts. However, prior works largely rely on static statistics aggregated across datasets, identifying heads that perform retrieval on average. This perspective overlooks the fine-grained temporal dynamics of autoregressive generation. In this paper, we investigate retrieval heads from a dynamic perspective. Through extensive analysis, we establish three core claims: (1) Dynamism: Retrieval heads vary dynamically across timesteps; (2) Irreplaceability: Dynamic retrieval heads are specific at each timestep and cannot be effectively replaced by static retrieval heads; and (3) Correlation: The model's hidden state encodes a predictive signal for future retrieval head patterns, indicating an internal planning mechanism. We validate these findings on the Needle-in-a-Haystack task and a multi-hop QA task, and quantify the differences on the utility of dynamic and static retrieval heads in a Dynamic Retrieval-Augmented Generation framework. Our study provides new insights into the internal mechanisms of LLMs.
Read the full abstract

All authors Yuping Lin, Zitao Li, Yue Xing, Pengfei He, Yingqian Cui, Yaliang Li, Bolin Ding, Jingren Zhou, Jiliang Tang

Computation and Language

Lowest Span Confidence: Zero-Shot Hallucination Detection from a Single LLM Response

Yitong Qiao, Licheng Pan, Yu Mi et al. (+6 more)

Hallucinations in Large Language Models (LLMs), i.e., plausible but non-factual generations, pose a significant challenge to reliable deployment in high-stakes environments. However, many existing hallucination detectors require expensive repeated sampling for consistency checks or access to model-internal states unavailable in common API-based scenarios. To this end, we propose an efficient zero-shot metric called Lowest Span Confidence (LSC) for hallucination detection under minimal resource assumptions. Concretely, LSC evaluates the local confidence of adjacent complete-word spans. By selecting the lowest aggregated confidence across neighboring words whose token widths can vary, LSC captures localized uncertainty associated with factual inconsistency. This boundary-aligned smoothing reduces the global dilution of perplexity and the sensitivity of minimum token probability to isolated noise. Our main evaluation spans four model families {Llama-2, Qwen2.5, Gemma-2, Mistral} and seven benchmarks {NQ, TriviaQA, SQuAD, CoQA, HotpotQA, RAGTruth, FELM}. Additional analyses examine word reconstruction, span width, and the role of adjacency in preserving local confidence. Across these settings, LSC is competitive with methods that use multiple responses or model-internal information while requiring only one response and its output token probabilities, without training a separate detector or using an auxiliary model.
Read the full abstract

All authors Yitong Qiao, Licheng Pan, Yu Mi, Lei Liu, Yue Shen, Jian Wang, Jinjie Gu, Fei Sun, Zhixuan Chu

Computation and Language

PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models

Haoyu Zheng, Yun Zhu, Yuqian Yuan et al. (+4 more)

Strategic planning is critical for multi-step reasoning, yet compact Large Language Models (LLMs) often lack the capacity to formulate global strategies, leading to error propagation in long-horizon tasks. Our analysis reveals that LLMs possess latent reasoning capabilities that can be unlocked when conditioned on explicit plans from a teacher model; however, runtime reliance on external guidance is often impractical due to latency and availability constraints. To bridge this gap, we propose PILOT (Planning via Internalized Latent Optimization Trajectories), a non-invasive framework designed to internalize the strategic oversight of large models into intrinsic Latent Guidance. Instead of altering backbone weights, PILOT employs a lightweight Hyper-Network to synthesize a query-conditioned Latent Guidance vector. This vector acts as an internal steering mechanism, guiding the model's representations toward optimal reasoning paths. Extensive experiments on mathematical and coding benchmarks demonstrate that PILOT effectively stabilizes reasoning trajectories, consistently outperforming strong baselines (e.g., +8.9% on MATH500) with negligible inference latency.
Read the full abstract

All authors Haoyu Zheng, Yun Zhu, Yuqian Yuan, Bo Yuan, Wenqiao Zhang, Siliang Tang, Jun Xiao

Computation and Language

PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review

Songjun Tu, Yiwen Ma, Jiahao Lin et al. (+6 more)

Large language models (LLMs) can generate fluent peer reviews but often miss sparse, subtle, and document-level scientific errors. We introduce PaperAudit-Bench, a benchmark for evaluating paper-level error detection and error-aware automated review. It comprises (1) PaperAudit-Dataset, 220 high-quality AI/ML papers containing more than 2,200 controlled synthetic errors across eight categories, and (2) PaperAudit-Review, an evidence-aware framework that links structured error detection to review generation. Experiments across models, error categories, paper sections, and analysis depths show that sparse long-context auditing remains challenging for current LLMs. Conditioning review generation on explicit, manuscript-grounded error evidence produces more conservative and discriminative assessments that align more closely with human judgments than representative automated-review baselines. The dataset also supports post-training compact detectors through SFT and RL; Qwen3-14B achieves 55.8% F1 at Detection@1 under a fixed held-out protocol. A PubPeer transfer study and deployment in ScienceOne, a scientific foundation-model platform led by the Chinese Academy of Sciences, further illustrate practical applicability. Code and Dataset Page: https://github.com/TU2021/PaperAudit ScienceOne Manuscript Review: https://www.scienceone.cn/lit/#/assistant
Read the full abstract

All authors Songjun Tu, Yiwen Ma, Jiahao Lin, Qichao Zhang, Xiangyuan Lan, Junfeng.Li, Nan Xu, Linjing Li, Dongbin Zhao

Computation and Language

Modeling Next-Token Prediction as Left-Nested Intuitionistic Implication

Paul Tarau

We introduce the Arrow Language Model, a neural architecture derived from an intuitionistic-logic interpretation of next-token prediction. Instead of representing tokens as additive embeddings mixed by attention, we encode a prefix as a left-nested implication chain whose structure preserves order through non-commutative composition. Next-token prediction corresponds to modus ponens, and sequence processing becomes constructive proof extension under the Curry–Howard correspondence. Our Prolog-based specialized theorem provers validate fundamental properties of the neural models, among which relations between commutative vs. non-commutative sequencing and single-token vs. multi-token prediction choices. We show that a neural architecture equivalent to multiplicative RNNs arises naturally from a proof-theoretic interpretation of next-token prediction as nested intuitionistic implication, we present a practical low-rank neural realization and position the model relative to Transformers and state-space models. Keywords: logic-based derivation of neural architectures, intuitionistic implicational logic, token-as-operator neural models, state-space models, alternatives to transformer-based foundational models.
Read the full abstract

Computation and Language · Cross-listed from eess.AS

The Voice of Equity: A Systematic Evaluation of Bias Mitigation Techniques for Speech-Based Cognitive Impairment Detection Across Architectures and Demographics

Yasaman Haghbin, Sina Rashidi, Ali Zolnour et al. (+1 more)

Speech-based detection of cognitive impairment offers a scalable, non-invasive screening, yet algorithmic bias across demographic and linguistic subgroups remains critically underexplored. We present the first comprehensive fairness analysis framework for speech-based multi-class cognitive impairment detection, systematically evaluating bias mitigation across architectures, and demographic subgroups. We developed two transformer-based architectures, SpeechCARE-AGF and Whisper-LWF-LoRA, on the multilingual NIA PREPARE Challenge dataset. Unlike prior work that typically examines single mitigation techniques, we compared pre-processing, in-processing, and post-processing approaches, assessing fairness via Equality of Opportunity and Equalized Odds across gender, age, education, and language. Both models achieved strong performance (F1: SpeechCARE-AGF 70.87, Whisper-LWF-LoRA 71.46) but exhibited substantial fairness disparities. Adults >=80 showed lower sensitivity versus younger groups; Spanish speakers demonstrated reduced TPR versus English speakers. Mitigation effectiveness varied by architecture: oversampling improved SpeechCARE-AGF for older adults (80+ TPR: 46.19%=>49.97%) but minimally affected Whisper-LWF-LoRA. This study addresses a critical healthcare AI gap by demonstrating that architectural design fundamentally shapes bias patterns and mitigation effectiveness. Adaptive fusion mechanisms enable flexible responses to data interventions, while frequency reweighting offers robust improvements across architectures. Our findings establish that fairness interventions must be tailored to both model architecture and demographic characteristics, providing a systematic framework for developing equitable speech-based screening tools essential for reducing diagnostic disparities in cognitive healthcare.
Read the full abstract

All authors Yasaman Haghbin, Sina Rashidi, Ali Zolnour, Maryam Zolnoori

Computation and Language

Towards Latent Diffusion Suitable For Text

Nesta Midavaine, Christian A. Naesseth and Grigory Bartosh

Language diffusion models aim to improve sampling speed and coherence over autoregressive LLMs. We introduce Neural Flow Diffusion Models for language generation, an extension of NFDM that enables the straightforward application of continuous diffusion models to discrete state spaces. NFDM learns a multivariate forward process from the data, ensuring that the forward process and generative trajectory are a good fit for language modeling. Our model substantially reduces the likelihood gap with autoregressive models of the same size, while achieving sample quality comparable to that of previous latent diffusion models.
Read the full abstract

Computation and Language

Intelligence Degradation in Long-Context LLMs: Critical Threshold Determination via Natural Length Distribution Analysis

Weiwei Wang, Jiyong Min and Weijie Zou

Large Language Models (LLMs) exhibit catastrophic performance degradation when processing contexts approaching certain critical thresholds, even when information remains relevant. This intelligence degradation-defined as over 30% drop in task performance-severely limits long-context applications. This degradation shows a common pattern: models maintain strong performance up to a critical threshold, then collapse catastrophically. We term this shallow long-context adaptation-models adapt for short to medium contexts but fail beyond critical thresholds. This paper presents three contributions: (1) Natural Length Distribution Analysis: We use each sample's natural token length without truncation or padding, providing stronger causal evidence that degradation results from context length itself. (2) Critical Threshold Determination: Through experiments on a mixed dataset (1,000 samples covering 5%-95% of context length), we identify the critical threshold for Qwen2.5-7B at 40-50% of maximum context length, where F1 scores drop from 0.55-0.56 to 0.3 (45.5% degradation), using five-method cross-validation. (3) Unified Framework: We consolidate shallow adaptation, explaining degradation patterns and providing a foundation for mitigation strategies. This work provides the first systematic characterization of intelligence degradation in open-source Qwen models, offering practical guidance for deploying LLMs in long-context scenarios.
Read the full abstract

Computation and Language

MALTopic: Multi-Agent LLM Topic Modeling Framework

Yash Sharma

Topic modeling is a crucial technique for extracting latent themes from unstructured text data, particularly valuable in analyzing survey responses. However, traditional methods often only consider free-text responses and do not natively incorporate structured or categorical survey responses for topic modeling. And they produce abstract topics, requiring extensive human interpretation. To address these limitations, we propose the Multi-Agent LLM Topic Modeling Framework (MALTopic). This framework decomposes topic modeling into specialized tasks executed by individual LLM agents: an enrichment agent leverages structured data to enhance textual responses, a topic modeling agent extracts latent themes, and a deduplication agent refines the results. Comparative analysis on a survey dataset demonstrates that MALTopic significantly improves topic coherence, diversity, and interpretability compared to LDA and BERTopic. By integrating structured data and employing a multi-agent approach, MALTopic generates human-readable topics with enhanced contextual relevance, offering a more effective solution for analyzing complex survey data.
Read the full abstract