Chain-of-thought (CoT) prompting is a de-facto standard technique to elicit reasoning-like responses from large language models (LLMs), allowing them to spell out individual steps before giving a final answer. While the resemblance to human-like reasoning is undeniable, the driving forces underpinning the success of CoT reasoning still remain largely unclear. In this work, we perform an in-depth analysis of CoT traces originating from competition-level mathematics questions, with the aim of better understanding how, and which parts of CoT actually contribute to the final answer. To this end, we introduce the notion of a potential, quantifying how much a given part of CoT increases the likelihood of a correct completion. Upon examination of reasoning traces through the lens of the potential, we identify surprising patterns including (1) its often strong non-monotonicity (due to reasoning tangents), (2) very sharp but sometimes tough to interpret spikes (reasoning insights and jumps) as well as (3) at times lucky guesses, where the model arrives at the correct answer without providing any relevant justifications before. While some of the behaviours of the potential are readily interpretable and align with human intuition (such as insights and tangents), others remain difficult to understand from a human perspective. To further quantify the reliance of LLMs on reasoning insights, we investigate the notion of CoT transferability, where we measure the potential of a weaker model under the partial CoT from another, stronger model. Indeed aligning with our previous results, we find that as little as 20% of partial CoT can ``unlock'' the performance of the weaker model on problems that were previously unsolvable for it, highlighting that a large part of the mechanics underpinning CoT are transferable.
Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces
University of Electronic Science and Technology of China（电子科技大学）；Singapore University of Technology and Design（新加坡科技设计大学）；Singapore Management University（新加坡管理学院）
TRACE: Toulmin-based Reasoning Assessment through Constructive Elements for LLM CoT Evaluation
Applied Agent Research Center, Korea Institute of Science（应用智能代理研究中心，韩国科学技术信息研究所）；Department of Computer Science and Engineering, University of Seoul, Republic of Korea（首尔大学计算机科学与工程系，大韩民国）
Detecting Unfaithful Chain-of-Thought via Circuit-Guided Internal-External Discrepancy
Jilin University（吉林大学）；University of Central Florida（中央佛罗里达大学）；Arizona State University（亚利桑那州立大学）；University of Vienna（维也纳大学）；University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）
The Challenge of Teaching Reasoning to LLMs Without RL or Distillation
Nvidia Corporation（Nvidia公司）；Institute for Artificial Intelligence Research（人工智能研究机构）；Faculty of Technical Sciences, University of Novi Sad（诺维萨德大学技术科学学院）；Department of Mathematics and Informatics, Faculty of Sciences, University of Novi Sad（诺维萨德大学科学学院数学与信息学系）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；University of Chicago（芝加哥大学）；Georgia Institute of Technology（佐治亚理工学院）；Department of Computer and Information Science, School of Engineering and Applied Science, University of Pennsylvania（宾夕法尼亚大学计算机与信息科学系）；Massachusetts Institute of Technology（麻省理工学院）
Are Reasoning LLMs Robust to Interventions on Their Chain-of-Thought?
Helmholtz Munich（海德堡-慕尼黑研究所）；Ludwig Maximilian University of Munich（慕尼黑路德维希-马克西米利安大学）；Technical University of Munich, Munich Center for Machine Learning (MCML)（慕尼黑技术大学，慕尼黑机器学习中心）
Select2Reason: Efficient Instruction-Tuning Data Selection for Long-CoT Reasoning
DataArc Tech Ltd.（DataArc科技有限公司）；Hong Kong University of Science and Technology (Guangzhou)（香港科学与技术大学（广州））；IDEA Research, International Digital Economy Academy（IDEA研究所，国际数字经济学院）；Hithink RoyalFlush Information Network Co., Ltd（Hithink RoyalFlush信息网络有限公司）
Focused Chain-of-Thought: Efficient LLM Reasoning via Structured Input Information
German Research Center for Artificial Intelligence (DFKI)（德国人工智能研究中心）；Technical University of Darmstadt（德累斯顿技术大学）；University of Kassel（卡塞尔大学）；TU Dortmund University（Dortmund技术大学）；TU Center for Trustworthy Data Science and Security, University Alliance Ruhr（TU可信数据科学与安全中心，鲁尔大学联盟）；Hessian Center for AI (Hessian.AI)（黑森人工智能中心（黑森.AI））；Centre for Cognitive Science, Technical University of Darmstadt（认知科学中心，德累斯顿技术大学）
SEER: Self-Enhancing Chain-of-Thought Compression for Reasoning Models
The State Key Laboratory of Blockchain and Data Security, Zhejiang University（区块链与数据安全国家重点实验室，浙江大学）；State Key Laboratory for Novel Software Technology, Nanjing University（新型软件技术国家重点实验室，南京大学）
Stable Reasoning, Unstable Responses: Mitigating LLM Deception via Stability Asymmetry
Institute for Artificial Intelligence, Peking University（北京大学人工智能研究院）；Beijing Academy of Artificial Intelligence（北京人工智能研究院）；School of Chinese as a Second Language, Peking University（北京大学第二语言学院）
CoT-X: An Adaptive Framework for Cross-Model Chain-of-Thought Transfer and Optimization
Purdue University（普渡大学）；Baruch College, City University of New York（纽约市立大学巴鲁克学院）；The Ohio State University（俄亥俄州立大学）；AI Agent Lab（AI Agent实验室）；Appcubic；Emory University（埃默里大学）
