跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+1030新增论文
+790其中 AI 相关
+127已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-11 21:33 · 全站最近入库：2026-11-11 18:36各信源进度 →

统计口径与数据新鲜度

统计 2026-11-10 21:58 — 2026-11-11 21:58（北京时间） · 点击数字查看对应简报；AI 相关包含在新增论文内。

最近完成检查：2026-11-11 21:33最近发现新增版本：2026-11-11 18:36库内论文最新更新时间：2026-11-11 08:00

每分钟刷新已保存的数据。空检查不代表采集异常；新收录可能包含历史论文。AI 导读按新增版本计数。

21,673 已收录论文18,321 AI 相关6,055 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

时间范围跟随排序：版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常，查看状态 →

找到 20 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-11 21:29（北京时间）
新论文 · v1arXiv:2611.14036

MISO-BR: An Open Corpus for Misogyny Detection in Brazilian Portuguese

Fábio A. S. Passos, Péricles B. C. Miranda, Ana Paula Cavalcanti

原文摘要

Misogyny is among the most prevalent forms of online hate speech directed at Brazilian users, yet open and auditable resources dedicated to it in Brazilian Portuguese remain scarce. We introduce MISO-BR, an open, documented and versioned corpus of 896 balanced social media texts annotated for binary misogyny, released with its construction protocol and reproducible baselines. The corpus combines lexicon- and target-guided scraping from X with items imported from OLID-BR, a large language model triage stage whose prompt and decoding parameters are released, and manual annotation by three annotators following a written guide. Inter-annotator reliability reaches a Fleiss' kappa of 0.788 over 816 items. Under a single fixed protocol we compare four classical classifiers (Multinomial and Bernoulli Naive Bayes, a linear SVM and a random forest) with fine-tuned BERTimbau Base and Large. The classical models cluster between 0.651 and 0.676 macro-F1 and are precision-oriented but recall-poor. The transformers reach 0.833 macro-F1 and accuracy (Wilson 95% CI for accuracy [0.77, 0.88]), a gap of roughly 16 points, while Base and Large are indistinguishable on the test set despite a 2.6x difference in training cost. We argue that the value of the resource lies less in its headline scores than in the auditability of its pipeline, and we state three limits that users must reckon with: the artificial class balance, whose effect on precision under realistic prevalence we quantify; the association between provenance and label, which the released source field makes measurable; and the selection bias induced by lexicon-guided sampling and automatic triage. The corpus, annotation guide, models and experimental scripts are publicly available.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.13968

Advancing a Global Framework for AI Safety and Governance for the Well-being of Humanity

Zeng Yi, Seán Ó hÉigeartaigh, Wang Zhengqi, Lu Enmeng, Chen Yu, Cao Gongce, Guo Xiaoyang, Kang Yanrong 等 76 位作者

原文摘要

The Report "Advancing a Global Framework for AI Safety and Governance for the Well-being of Humanity" is an important outcome of the AI Safety and Governance Program of World Internet Conference Specialized Committee on Artificial Intelligence in 2025. This report provides a systematic review of current global explorations and practices in AI safety and governance, identifies key issues that need to be addressed, and draws upon multilateral governance experiences from other domains. Centered on mechanism design, it explores how to build a global AI safety and governance system that ensures safety and security, promotes inclusiveness, clarifies responsibilities and accountabilities, strengthens effective coordination, and ensures authoritative and efficient governance. Guided by the principles of multilateralism and the vision of building a community with a shared future for humankind, the report proposes institutional measures and policy recommendations to establish a United Nations-centered global AI safety and governance framework. It aims to serve as a reference for governments, international organizations, enterprises, and research institutions. The report is to contribute insights and momentum to advancing people-centered and beneficial AI governance worldwide, ensuring that AI technologies truly serve the well-being of all humanity.

cs.CY

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2611.05355

Reading Why: Born-Measured Attention with Exact Word Evidence on Light and Qubits

Teerawat Tongloy, Jiraporn Pongsopa, Rangsan Jomtarak

原文摘要

When a language model answers, we can usually only estimate why. We build a transformer in which every nonlinearity is a quadratic detection (an interference, a balanced interferometric reading or a camera) and the output is a Born probability: attention $|\langle q|k\rangle|^2$, the form of the Born rule, normalised over each row, a measured product in place of the feed-forward nonlinearity, and a Born readout; between detections a classical controller squares, divides, rescales and re-encodes. With its denominators held at their values for the input, each class probability is a polynomial in the word amplitudes, so its integrated-gradient evidence is computed by a finite quadrature with no quadrature error: the words' contributions sum to the class probability above its no-information share, to $10^{-13}$ or better. On e-SNLI natural language inference the model reaches 77.45 +- 0.91 % test accuracy, level within the spread of three seeds with a softmax transformer of the same size (76.62 +- 0.36 %). Deleting the word its evidence ranks first changes the answer in 44 % of test pairs, out of the 53 % any single deletion can change, against 32 % for the softmax model's integrated gradients and 31 % for the Born Transformer's own attention. The same weights compile exactly to optoelectronic optics (simulated), and its attention overlaps to Hadamard tests on qubits; a separate coherent attention layer becomes one post-selected circuit (852 CNOTs on 23 qubits). On IBM's ibm_kingston, reading only the attention overlaps, the device gave the model's answers on the two shortest test pairs, also when the one word that decides one of them was deleted; with calibration circuits as deep as the overlaps, each word's evidence with the device's attention row sums followed the exact evidence on one pair (correlation 0.997) and on the other less closely (0.962).

quant-ph · cs.CL · cs.LG · physics.optics

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1104ACL Anthology · ACL 2026

Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment

Bobo Li National University of Singapore, Rui Wu Sichuan University, Zibo Ji, Meishan Zhang Harbin Institute of Technology (Shenzhen), China, Hao Fei National University of Singapore and University of Oxford, Min Zhang Harbin Institute of Technology, Shenzhen and Harbin Institute of Technology, Mong-Li Lee National University of Singapore, Wynne Hsu National University of Singapore

原文摘要

Large Language Model agents have rapidly evolved from static text generators into dynamic systems capable of executing complex autonomous workflows. To enhance reliability, multi-agent frameworks assigning specialized roles are increasingly adopted to enable self-reflection and mutual auditing. While such role-playing effectively leverages domain expert knowledge, we find it simultaneously induces a human-like cognitive bias known as Actor-Observer Asymmetry (AOA). Specifically, an agent acting as an actor (during self-reflection) tends to attribute failures to external factors, whereas an observer (during mutual auditing) attributes the same errors to internal faults. We quantify this using our new Ambiguous Failure Benchmark, which reveals that simply swapping perspectives triggers the AOA effect in over 20% of cases for most models. To tame this bias, we introduce ReTAS (Reasoning via Thesis-Antithesis-Synthesis), a model trained through dialectical alignment to enforce perspective-invariant reasoning. By integrating dialectical chain-of-thought with Group Relative Policy Optimization, ReTAS guides agents to synthesize conflicting viewpoints into an objective consensus. Experiments demonstrate that ReTAS effectively mitigates attribution inconsistency and significantly improves fault resolution rates in ambiguous scenarios.

cs.CL · cs.MA

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.131ACL Anthology · ACL 2026

RealTalk-CN: A Realistic Chinese Speech Task-Oriented Dialogue Benchmark with Cross-Modal Analysis

Enzhi Wang, Jiaming Zhou, Yuhang Jia, Aobo Kong, Qicheng Li Nankai University, Yong Qin Nankai University

原文摘要

Recent advances in speech large language models (e.g., GPT-4o) have enabled end-to-end spoken interactions, yet their robustness in real-world applications remains unclear, where systems must assist users in completing specific tasks under complex conditions such as multi-turn, ambiguous, and often spontaneous speech, as well as natural alternation between speech and text. Task-oriented dialogue (TOD) offers a realistic scenario to evaluate whether models can effectively help users accomplish such task-oriented goals, but existing benchmarks are mainly text-based, and the few speech datasets are limited to English and often neglect spontaneous disfluencies and speaker diversity. To address this gap, we introduce RealTalk-CN, the first Chinese multi-turn, multi-domain speech–text TOD dataset, containing 5.4k dialogues (60K turns, ~150 hours) of real human-to-human recordings with detailed annotations for dialogue states, disfluency types, and speaker characteristics. Based on this dataset, we propose a cross-modal interaction task supporting dynamic speech-text switching and a comprehensive evaluation protocol assessing robustness to disfluencies, sensitivity to speaker variation, and cross-domain generalization. Experiments on state-of-the-art models demonstrate the challenges posed by RealTalk-CN and establish its value as a benchmark for developing reliable and fair Speech LLMs in real-world deployments. The dataset and evaluation framework are available to encourage further research.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1665ACL Anthology · ACL 2026

DUD: Decoupled Update Dynamics for Reliable Uncertainty Quantification in Large Language Models

Yixin Bu, Runze Xia, Guanyun Zou Nanjing University of Aeronautics and Astronautics, Yupeng Ji, Hongliang Dai Nanjing University of Aeronautics and Astronautics, Haodong Liu, Piji Li Nanjing University of Aeronautics and Astronautics

原文摘要

Accurate Uncertainty Quantification (UQ) is critical for reliable deployment of Large Language Models (LLMs), yet traditional probability-based metrics often fail to capture the model’s true epistemic state. While recent mechanistic approaches leverage hidden state dynamics, they typically aggregate residual stream updates, conflating the distinct roles of parametric memory (Feed-Forward Networks) and contextual processing (Attention). We argue that this aggregation obscures fine-grained mechanistic conflicts, such as memory-context misalignment, that are fundamental indicators of uncertainty. To address this, we introduce Decoupled Update Dynamics (DUD), a framework that explicitly decouples FFN and Attention contributions via noise-induced causal interventions. By quantifying the independent restoration capabilities of each module, we construct a dual-stream dynamic profile that captures the model’s internal fragility. Extensive experiments demonstrate that DUD significantly outperforms state-of-the-art baselines in both uncertainty estimation and calibration, while exhibiting superior cross-dataset generalization, validating decoupled dynamics as a robust proxy for model faithfulness.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1759ACL Anthology · ACL 2026

Mitigating Over-Refusal in Aligned Large Language Models via Inference-Time Activation Energy

Eric Hanchen Jiang, Weixuan Ou, Run Liu, Shengyuan Pang, Guancheng Wan WanFlow AI and Wuhan University, Ranjie Duan Tencent, Wei Dong Nanyang Technological University, Kai-Wei Chang University of California, Los Angeles and Amazon 等 11 位作者

原文摘要

Safety alignment of large language models currently faces a central challenge: existing alignment techniques often prioritize mitigating responses to harmful prompts at the expense of overcautious behavior, leading models to incorrectly refuse benign requests. A key goal of safe alignment is therefore to improve safety while simultaneously minimizing false refusals. In this work, we introduce Energy Landscape Steering (ELS), a novel, fine-tuning free framework designed to resolve this challenge through dynamic, inference-time intervention. We trained a lightweight, external Energy-Based Model (EBM) to assign high energy to undesirable (false refusal or jailbreak) states and low energy to desirable (helpful response or safe reject) ones. During inference, the EBM maps the LLM’s internal activations to an energy landscape, and we use the gradient of the energy function to steer the hidden states toward low-energy regions in real time. This dynamically guides the model toward desirable behavior without modifying its parameters. By decoupling behavioral control from the model’s core knowledge, ELS provides a flexible and computationally efficient solution. Extensive experiments across diverse models demonstrate its effectiveness: raising compliance on the ORB-H benchmark from 57.3% to 82.6% while maintaining the baseline safety performance. Our work establishes a promising paradigm for building LLMs that simultaneously achieve high safety and low false refusal rates.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1764ACL Anthology · ACL 2026

Dynamic Generation of Multi LLM Agents Communication Topologies with Graph Diffusion Models

Eric Hanchen Jiang, Levina Li, Frank Wan, Xiao Liang, Sophia Yin, Yuchen Wu University of Washington, Xinfeng Li Nanyang Technological University, Yizhou Sun University of California, Los Angeles and Amazon 等 11 位作者

原文摘要

The efficiency of multi-agent systems driven by large language models (LLMs) largely hinges on their communication topology. However, designing an optimal topology is a non-trivial challenge, as it requires balancing competing objectives such as task performance, communication cost, and robustness. Existing frameworks often rely on static or hand-crafted topologies, which inherently fail to adapt to diverse task requirements, leading to either excessive token consumption for simple problems or performance bottlenecks for complex ones. To address this challenge, we introduce a novel generative framework called Guided Topology Diffusion (GTD). Inspired by conditional discrete graph diffusion models, GTD formulates topology synthesis as an iterative construction process. At each step, the generation is steered by a lightweight proxy model that predicts multi-objective rewards (e.g., accuracy, utility, cost), enabling real-time, gradient-free optimization towards task-adaptive topologies. This iterative, guided synthesis process distinguishes GTD from single-step generative frameworks, enabling it to better navigate complex design trade-offs. We validated GTD across multiple benchmarks, and experiments show that this framework can generate highly task-adaptive, sparse, and efficient communication topologies, significantly outperforming existing methods in LLM agent collaboration. Our code is available at https://anonymous.4open.science/r/diffusion_agent-953C.

cs.CL · cs.MA

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.180ACL Anthology · ACL 2026

DMHM: Density-aware Manifold Learning and Hybrid Mahalanobis Energy for LLMs-generated Text Detection

Tianle Liu, Zhiliang Tian National University of Defense Technology, Zhen Huang National University of Defense Technology, Tianlun Liu National University of Defense Technology, Jingyuan Huang, Zhaoning Zhang National University of Defense Technology, Chengcheng Shao, Dongsheng Li National University of Defense Technology

原文摘要

As the text generated by large language models (LLMs) increasingly resembles human-written text (HWT), detecting LLM-generated text (LGT) is crucial to avoid malicious use of LGT. Recent research treats LGT detection as an out-of-distribution (OOD) detection problem and views HWT as the OOD. However, existing OOD detection methods assume that LGT is a single homogeneous distribution. In practice, LGT exhibits different characteristics under different generation conditions. Text from weaker LLMs tends to form distinct clusters and is easy to detect, whereas text from stronger models significantly overlaps with HWTs and is hard to detect. To address the issue, in this paper, we propose an LGT detection framework based on density-aware manifold learning and the construction of hybrid Mahalanobis energy. We apply density-aware manifold learning with Laplacian smoothness and density regularization in embedding space, amplifying differences between LGT and HWT. We further propose a density-adaptive hybrid Mahalanobis metric that combines global and local covariance via density weighting, enabling adaptation to the manifold-aware embedding space. Finally, based on the metric, we define the distribution energy as a measure of distribution discrepancy, and we employ energy learning and contrastive learning to separate distributions hierarchically, establishing a clear OOD decision boundary. Experiments show that our method outperforms strong baselines.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1824ACL Anthology · ACL 2026

Knowledge Injection Exists in MoE? Exploring Expert-Aware Contrast Decoding in MoE for Mitigating LLMs’ Hallucinations

Xinyue Fang, Zhiliang Tian National University of Defense Technology, Zhen Huang National University of Defense Technology, Ziyi Pan, Zhihua Wen, Xi Wang, Quntian Fang, Dongsheng Li National University of Defense Technology

原文摘要

Existing LLM hallucination mitigation methods, including prompt engineering and model optimization, either hardly alter models’ internal knowledge or have poor cross-domain generalization. Contrastive decoding mitigates hallucinations by using layer-wise differences in LLMs. However, prior studies only explore transformer-based models (e.g., GPT), ignoring other effective frameworks like mixture-of-experts (MoE) models. Since MoE alters the traditional transformer architecture, we conduct empirical studies to investigate whether similar layer-wise differences exist in MoEs. Our results show that they do not exist in MoE with shared experts; nevertheless, across different MoEs, higher layers exhibit distinct expert activation patterns between factual and non-factual outputs. Building on these, we propose EAACD, an expert-aware adaptive contrast decoding that uses expert differences in MoE’s higher layers to mitigate hallucinations on QA tasks. EAACD splits high-layer experts into a higher-reliability group and several lower-reliability groups based on their confidence and consistency. It contrasts the higher-reliability group’s prediction with each lower-reliability group’s prediction to calibrate the model’s original predictions. To strengthen this contrast, EAACD amplifies hallucinations from lower-reliability experts via attention and masking to provide stronger negative references. EAACD outperforms all baselines on four datasets

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.187ACL Anthology · ACL 2026

CTTA-T: Continual Test-Time Adaptation for Text Understanding via Teacher-Student with a Domain-aware and Generalized Teacher

Tianlun Liu National University of Defense Technology, Zhiliang Tian National University of Defense Technology, Zhen Huang National University of Defense Technology, Xingzhi Zhou Huawei Technologies Ltd., Wanlong Yu, Tianle Liu, Feng Liu National University of Defense Technology, Dongsheng Li National University of Defense Technology

原文摘要

Text understanding application often suffers from domain shifts. To handle testing domains, domain adaptation (DA) is trained to adapt to a fixed and observed testing domain; a more challenging paradigm, test-time adaptation (TTA), cannot access the testing domain during training and online adapts to the testing samples during testing, where the samples are from a fixed domain. We aim to explore a more practical and underexplored scenario, continual test-time adaptation (CTTA) for text understanding, which involves a sequence of testing (unobserved) domains in testing. Current CTTA methods struggle in reducing error accumulation over domains and enhancing generalization to handle unobserved domains: 1) Noise-filtering reduces accumulated errors but discards useful information, and 2) accumulating historical domains enhances generalization, but it is hard to achieve adaptive accumulation. In this paper, we propose a CTTA-T (continual test-time adaptation for text understanding) framework adaptable to evolving target domains: CTTA-T adopts a teacher-student framework, where the teacher is equipped with domain awareness and generalization for evolving domains. To improve teacher predictions, we propose a refine-then-filter based on dropout-driven consistency, which calibrates predictions and removes unreliable guidance. For the adaptation–generalization trade-off, we construct a domain-aware teacher by dynamically accumulating cross-domain semantics via incremental PCA, which continuously tracks domain shifts. Experiments show CTTA-T excels baselines.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.200ACL Anthology · ACL 2026

Thermometer of Thoughts: Enhancing LLM’s Exploration via Attention Temperature Modulation

Zhiyuan Yu, Shijian Xiao, Cam-Tu Nguyen, Zhangyue Yin, Lekai Xing, Wenzhong Li Nanjing University, Sanglu Lu Nanjing University

原文摘要

Improving the exploration of reasoning is essential for advancing Large Language Models’ (LLMs) problem-solving performance. Current methods primarily rely on output-level stochasticity, which decode within fixed reasoning patterns of LLM and suffer from insufficient exploration. In this paper, we introduce adjusting attention temperature to directly modulate the model’s internal focus during reasoning, which enables a dynamic shift between exploratory and focused processing. We reveal that moderate adjustments preserve LLM’s reasoning capability while producing problem hardness-dependent benefits: higher temperatures facilitate solving complex tasks by encouraging wider exploration, whereas lower temperatures mitigate overthinking on simpler problems. Leveraging this insight, we propose a two-stage inference strategy: first, attention temperature scaling modulates the LLM’s reasoning patterns to diversify the reasoning traces; then, a difficulty-aware aggregation scheme is introduced to effectively identify the most reliable solution from the generated candidates. Extensive evaluations show that our method improves Pass@10 by 6.78–14.20% and aggregation accuracy by 9.74% across 7 reasoning benchmarks.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.2098ACL Anthology · ACL 2026

Legal Judgment Prediction: A Reflection on the State of the Art

Yi Feng Nanjing University, Chuanyi Li nanjing university, Vincent Ng

原文摘要

Automatic legal judgment prediction (LJP) has recently received increasing attention in the natural language processing community because of its practical values in the real world. Significant progress has been achieved on LJP in the past decade. However, most existing LJP research primarily focuses on developing methods that achieve better performance on standard evaluation datasets, with limited emphasis on the long-term advancement of the field beyond improving evaluation metrics. In this position paper, we reflect on the state of the art in LJP research, and explore issues that should motivate researchers to think beyond merely enhancing performance metrics, with the ultimate goal of sparking discussions among LJP researchers about the future trajectory of the field.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.348ACL Anthology · ACL 2026

JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification

Xi Wang, Songlei Jian National University of Defense Technology, Shasha Li National University of Defense Technology, Xiaopeng Li National University of Defense Technology, Zhaoye Li, Bin Ji National University of Defense Technology, Baosheng Wang, Jie Yu

原文摘要

Despite extensive safety alignment, Large Language Models (LLMs) often fail against jailbreak attacks. While machine unlearning has emerged as a promising defense by erasing specific harmful parameters, current methods remain vulnerable to diverse jailbreaks. We first conduct an empirical study and discover that this failure mechanism is caused by jailbreaks primarily activating non-erased parameters in the intermediate layers. Further, by probing the underlying mechanism through which these circumvented parameters reassemble into the prohibited output, we verify the persistent existence of dynamic jailbreak paths and show that the inability to rectify them constitutes the fundamental gap in existing unlearning defenses. To bridge this gap, we propose Jailbreak Path Unlearning (JPU), which is the first to rectify dynamic jailbreak paths towards safety anchors by dynamically mining on-policy adversarial samples to expose vulnerabilities and identify jailbreak paths. Extensive experiments demonstrate that JPU significantly enhances jailbreak resistance against dynamic attacks while preserving the model’s utility.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.376ACL Anthology · ACL 2026

The Reasoning Trap: How Enhancing LLM Reasoning Amplifies Tool Hallucination

Chenlong Yin Pennsylvania State University, Zeyang Sha Ant Group, Shiwen Cui ant group, Changhua Meng Ant Group, Zechao Li Nanjing University of Science and Techonolgy

原文摘要

Enhancing the reasoning capabilities of Large Language Models (LLMs) is a key strategy for building agents that ”think then act”. However, recent observations, like OpenAI’s o3, suggest a paradox: stronger reasoning often coincides with increased hallucination, yet no prior work has systematically examined whether reasoning enhancement itself causes tool hallucination. We address this gap with the central question: Does strengthening reasoning increase tool hallucination? To answer this, we introduce SimpleToolHalluBench, a diagnostic benchmark measuring tool hallucination in two failure modes: (i) no tool available, and (ii) only distractor tools available. Through controlled experiments, we establish three key findings. First, we demonstrate a causal relationship: progressively enhancing reasoning through RL increases tool hallucination proportionally with task performance gains. Second, this effect transcends overfitting—training on non-tool tasks (e.g., mathematics) still amplifies subsequent tool hallucination. Third, the effect is method-agnostic, appearing when reasoning is instilled via supervised fine-tuning and when it is merely elicited at inference by switching from direct answers to step-by-step thinking. We also evaluate mitigation strategies including Prompt Engineering and Direct Preference Optimization (DPO), revealing a fundamental reliability–capability trade-off: reducing hallucination consistently degrades utility. Mechanistically, Reasoning RL disproportionately collapses tool-reliability–related representations, and hallucinations surface as amplified divergences concentrated in late-layer residual streams. These findings reveal that current reasoning enhancement methods inherently amplify tool hallucination, highlighting the need for new training objectives that jointly optimize for capability and reliability. Our implementation is provided at https://github.com/albert-y1n/Reasoning_Trap.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.437ACL Anthology · ACL 2026

Alloc-MoE: Budget-Aware Expert Activation Allocation for Efficient Mixture-of-Experts Inference

Baihui Liu National University of Defense Technology, Kaiyuan Tian National University of Defense Technology, Wei Wang National University of Defense Technology, Zhaoning Zhang National University of Defense Technology, Linbo Qiao, Dongsheng Li National University of Defense Technology

原文摘要

Mixture-of-Experts (MoE) has become a dominant architecture for scaling large language models due to their sparse activation mechanism. However, the substantial number of expert activations creates a critical latency bottleneck during inference, especially in resource-constrained deployment scenarios. Existing approaches that reduce expert activations potentially lead to severe model performance degradation. In this work, we introduce the concept of activation budget as a constraint on the number of expert activations and propose Alloc-MoE, a unified framework that optimizes budget allocation coordinately at both the layer and token levels to minimize performance degradation. At the layer level, we introduce Alloc-L, which leverages sensitivity profiling and dynamic programming to determine the optimal allocation of expert activations across layers. At the token level, we propose Alloc-T, which dynamically redistributes activations based on routing scores, optimizing budget allocation without increasing latency. Extensive experiments across multiple MoE models demonstrate that Alloc-MoE maintains model performance under a constrained activation budget. Especially, Alloc-MoE achieves 1.15\times prefill and 1.34\times decode speedups on DeepSeek-V2-Lite at half of the original budget.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.450ACL Anthology · ACL 2026

Sparrow: Text-Anchored Window Attention with Visual-Semantic Glimpsing for Speculative Decoding in Video LLMs

Libo Zhang, Zhaoning Zhang National University of Defense Technology, Hongwanyang, Peng Qiao, Dongsheng Li National University of Defense Technology

原文摘要

Although speculative decoding is widely used to accelerate Vision-Language Models (VLMs) inference, it faces severe performance collapse when applied to Video Large Language Models (Vid-LLMs). The draft model typically falls into the trap of attention dilution and negative visual gain due to key-value cache explosion and context window mismatches. We observe a visual semantic internalization phenomenon in Vid-LLMs, indicating that critical visual semantics are implicitly encoded into text hidden states during deep-layer interactions, which renders raw visual inputs structurally redundant during deep inference. To address this, we propose the Sparrow framework, which first utilizes visually-aware text-anchored window attention via hidden state reuse to fully offload visual computation to the target model, and leverages intermediate-layer visual state bridging to train the draft model with semantic-rich intermediate states, thereby filtering out low-level visual noise. Additionally, a multi-token prediction strategy is introduced to bridge the training-inference distribution shift. Experiments show that Sparrow achieves an average speedup of 2.82x even with 25k visual tokens, effectively resolving the performance degradation in long sequences and offering a practical solution for real-time long video tasks.

cs.CL · cs.CV

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.455ACL Anthology · ACL 2026

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning

Biao Yi, Tiansheng Huang, Baolei Zhang Nankai University, Tong Li Nankai University, Lihai Nie, Zheli Liu Nankai University, Li Shen Sun Yat-Sen University

原文摘要

Fine-tuning-as-a-service, while commercially successful for Large Language Model (LLM) providers, exposes models to harmful fine-tuning attacks. As a widely explored defense paradigm against such attacks, unlearning attempts to remove malicious knowledge from LLMs, thereby essentially preventing them from being used to perform malicious tasks. However, we highlight a critical flaw: the inherent general adaptability of LLMs allows them to easily bypass selective unlearning by rapidly relearning or repurposing their general capabilities for harmful tasks. To address this fundamental limitation, we propose a paradigm shift: instead of selective removal, we advocate for inducing model collapse, effectively forcing the model to ”unlearn everything”, specifically in response to updates characteristic of malicious adaptation. This collapse directly neutralizes the very general capabilities that attackers exploit, tackling the core issue unaddressed by selective unlearning. We introduce the Collapse Trap (CTRAP) as a practical mechanism to implement this concept conditionally. Embedded during alignment, CTRAP pre-configures the model’s reaction to subsequent fine-tuning dynamics. If updates during fine-tuning constitute a persistent attempt to reverse safety alignment, the pre-configured trap triggers a progressive degradation of the model’s core language modeling abilities, ultimately rendering it inert and useless for the attacker. Crucially, this collapse mechanism remains dormant during benign fine-tuning, ensuring the model’s utility and general capabilities are preserved.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.850ACL Anthology · ACL 2026

MeepleLM: A Virtual Playtester Simulating Diverse Subjective Experiences

Zizhen Li Nankai University, Chuanhao Li, Yibin Wang, Jianwen Sun Nankai University, Yukang Feng, Jiaxin Ai Wuhan University, Fanrui Zhang, Mingzhu Sun Nankai University 等 10 位作者

原文摘要

Recent advancements have expanded the role of Large Language Models (LLMs) in board games from playing agents to creative co-designers. However, a critical gap remains: current systems lack the capacity to offer constructive critique grounded in the emergent user experience. Bridging this gap is fundamental for harmonizing Human-AI collaboration, as it empowers designers to refine their creations via external perspectives while steering models away from biased or unpredictable outcomes. Automating this evaluation presents two challenges: inferring the latent dynamics connecting static rules to gameplay without an explicit engine, and modeling the subjective heterogeneity of diverse player groups. To address these, we curate a comprehensive dataset of 1,727 structurally corrected rulebooks and 150K reviews selected via rigorous quality scoring and facet-aware sampling. We augment this data with Mechanics-Dynamics-Aesthetics (MDA) reasoning to explicitly bridge the causal gap between written rules and player experience. We further distill distinct player personas and introduce MeepleLM, a specialized model that internalizes persona-specific reasoning patterns to accurately simulate the subjective feedback of diverse player archetypes. Extensive experiments demonstrate that MeepleLM significantly outperforms latest commercial models (e.g., GPT-5.1, Gemini3-Pro) in community alignment and critique quality, achieving a 70% preference rate in user studies assessing practical utility. MeepleLM serves as a reliable virtual playtester that provides experience-grounded feedback, offering a practical step towards audience-aligned Human-AI collaboration.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.949ACL Anthology · ACL 2026

Language Reconstruction with Brain Predictive Coding from fMRI Data

Congchi Yin Nanjing University of Aeronautics and Astronautics, Ziyi Ye Fudan University, Piji Li Nanjing University of Aeronautics and Astronautics

原文摘要

Many recent studies have shown that the perception of speech can be decoded from brain signals and subsequently reconstructed as continuous language. However, there is a lack of neurological basis for how the semantic information embedded within brain signals can be used more effectively to guide language reconstruction. Predictive coding theory suggests the human brain naturally engages in continuously predicting future words that span multiple timescales. This implies that the decoding of brain signals could potentially be associated with a predictable future. To explore the predictive coding theory within the context of language reconstruction, this paper proposes PredFT (FMRI-to-Text decoding with Predictive coding). PredFT consists of a main network and a side network. The side network obtains brain predictive representation from related regions of interest (ROIs) with a self-attention module. The representation is then fused into the main network for continuous language decoding. Experiments on two naturalistic language comprehension fMRI datasets show that PredFT outperforms current decoding models on several evaluation metrics.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表