Dissecting Failure Dynamics in Large Language Model Reasoning
School of Information Science and Engineering, Yunnan University, Kunming, China(云南大学信息科学与工程学院,昆明,中国);Yunnan Key Laboratory of Intelligent Systems and Computing, Kunming, China(云南省智能系统与计算重点实验室,昆明,中国)
Wei Zhu, Jian Zhang, Lixing Yu, Kun Yue, Zhiwen Tang
Large Language Models (LLMs) achieve strong performance through extended inference-time deliberation, yet how their reasoning failures arise remains poorly understood. By analyzing model-generated reasoning trajectories, we find that errors are not uniformly distributed but often originate from a small number of early transition points, after which reasoning remains locally coherent but globally incorrect. These transitions coincide with localized spikes in token-level entropy, and alternative continuations from the same intermediate state can still lead to correct solutions. Based on these observations, we introduce GUARD, a targeted inference-time framework that probes and redirects critical transitions using uncertainty signals. Empirical evaluations across multiple benchmarks confirm that interventions guided by these failure dynamics lead to more reliable reasoning outcomes. Our findings highlight the importance of understanding when and how reasoning first deviates, complementing existing approaches that focus on scaling inference-time computation.
Reasoning that Travels: Dissecting How Chain-of-Thought Transfers Across Models
MaiNLP, Center for Information and Language Processing, LMU Munich, Germany(MaiNLP,信息与语言处理中心,慕尼黑大学,德国);Munich Center for Machine Learning, Germany(慕尼黑机器学习中心,德国)
SafeRBench: Dissecting the Reasoning Safety of Large Language Models
Nanjing University(南京大学);Fudan University(复旦大学);Carnegie Mellon University(卡内基梅隆大学);Chinese Academy of Sciences(中国科学院);Nanyang Technological University(南洋理工大学)
Dissecting Role Cognition in Medical LLMs via Neuronal Ablation
School of Information, Renmin University of China(中国人民大学信息学院);Center of Machine Learning Research, Peking University(北京大学机器学习研究中心);School of Artificial Intelligence, Shanghai Jiaotong University(上海交通大学人工智能学院);Health Informatics Committee of the China Society for Scientific and Technical Information, Renmin University of China(中国科学技术信息学会健康信息技术委员会,中国人民大学);Institute for Advanced Algorithms Research (IAAR), Shanghai, China(先进算法研究所(IAAR),上海,中国);Memtensor Research Center, Shanghai, China(Memtensor研究中心,上海,中国)
Dissecting AI Trading: Behavioral Finance and Market Bubbles
Saïd Business School, University of Oxford(牛津大学said商学院);School of Management and Economics, The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)管理学院)
Systematic Failures in Collective Reasoning under Distributed Information in Multi-Agent LLMs
School of Computer Science, Carnegie Mellon University, Pittsburgh, USA(计算机科学学院,卡内基梅隆大学,匹兹堡,美国);School of Environment and Society, Institute of Science Tokyo, Tokyo, Japan(环境与社会学院,东京科学研究所,东京,日本)
The Periodic Table of LLM Reasoning: A Structured Survey of Reasoning Paradigms, Methods, and Failure Modes
Singapore Institute of Technology(新加坡理工大学);Nvidia AI Center (SNAIC)(英伟达人工智能中心(SNAIC));MIDAS Lab, IIIT Delhi(IIIT德里MIDAS实验室);MIDAS Lab, IIT Mandi(IIT曼迪MIDAS实验室);Owl Autonomous Imaging, Inc.(Owl自主成像公司);College of Computing & Data Science, NTU Singapore(新加坡南洋理工大学计算与数据科学学院);NVIDIA AI Technology Centre, Singapore(英伟达新加坡人工智能技术中心);Department of Computer Science and Engineering, IIT Kanpur(IIT坎普尔计算机科学与工程系)
CogniLoad: A Synthetic Natural Language Reasoning Benchmark With Tunable Length, Intrinsic Difficulty, and Distractor Density
UiT - The Arctic University of Norway(UiT - 北极大学);Integreat - Norwegian Centre for knowledge-driven machine learning(Integreat - 挪威知识驱动机器学习中心);University of Oslo(奥斯陆大学)
Catching The Correct Answer Trap: Characterising AI Tutor Blind Spots When Analysing Student Reasoning
Department of Computer Science, University College London, UK(英国伦敦大学学院计算机科学系);Centre for Artificial Intelligence, University College London, UK(英国伦敦大学学院人工智能中心)
Two Failures of Self-Consistency in the Multi-Step Reasoning of LLMs
Center for Data Science, New York University(纽约大学数据科学中心);New York University(纽约大学);Department of Linguistics, New York University(纽约大学语言学系);Google(谷歌公司);NYU Shanghai(上海纽约大学);Anthropic(Anthropic公司)