Supervised Fine-Tuning (SFT) on long Chain-of-Thought (CoT) trajectories has become a pivotal phase in building large reasoning models. However, how CoT trajectories from different sources influence the generalization performance of models remains an open question. In this paper, we conduct a comparative study using two sources of verified CoT trajectories generated by two competing models, \texttt{DeepSeek-R1-0528} and \texttt{gpt-oss-120b}, with their problem sets controlled to be identical. Despite their comparable performance, we uncover a striking paradox: lower training loss does not translate to better generalization. SFT on \texttt{DeepSeek-R1-0528} data achieves remarkably lower training loss, yet exhibits significantly worse generalization performance on reasoning benchmarks compared to those trained on \texttt{gpt-oss-120b}. To understand this paradox, we perform a multi-faceted analysis probing token-level SFT loss and step-level reasoning behaviors. Our analysis reveals a difference in reasoning patterns. \texttt{gpt-oss-120b} exhibits highly convergent and deductive trajectories, whereas \texttt{DeepSeek-R1-0528} favors a divergent and branch-heavy exploration pattern. Consequently, models trained with \texttt{DeepSeek-R1} data inherit inefficient exploration behaviors, often getting trapped in redundant exploratory branches that hinder them from reaching correct solutions. Building upon this insight, we propose a simple yet effective remedy of filtering out frequently branching trajectories to improve the generalization of SFT. Experiments show that training on selected \texttt{DeepSeek-R1-0528} subsets surprisingly improves reasoning performance by up to 5.1% on AIME25, 5.5% on BeyondAIME, and on average 3.6% on five benchmarks.
RefCritic: Training Long Chain-of-Thought Critic Models with Refinement Feedback
Chinese Information Processing Laboratory, Institute of Software, Chinese Academy of Sciences(中国科学院软件研究所信息处理实验室);Alibaba Group(阿里巴巴集团);University of Chinese Academy of Sciences(中国科学院大学)
Symbolic Chain-of-Thought Distillation: Small Models Can Also "Think" Step-by-Step
University of California Los Angeles(加利福尼亚大学洛杉矶分校);Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of Southern California(南加利福尼亚大学);Yonsei University(延世大学);University of Washington(华盛顿大学)
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)
Brain-Inspired Two-Stage Approach: Enhancing Mathematical Reasoning by Imitating Human Thought Processes
School of Information Science and Technology, ShanghaiTech University(上海科技大学信息科学与技术学院);Shanghai Innovation Center for Processor Technologies(上海处理器技术创新中心);School of Information Science and Technology, University of Science and Technology of China(中国科学技术大学信息科学与技术学院)
Cognitive Chain-of-Thought (CoCoT): Structured Multimodal Reasoning about Social Situations
Seoul National University(首尔国立大学);Human-Computer Interaction Institute, Carnegie Mellon University(人机交互研究所,卡内基梅隆大学);Language Technologies Institute, Carnegie Mellon University(语言技术研究所,卡内基梅隆大学)
Chaining Simultaneous Thoughts for Numerical Reasoning
Tsinghua University(清华大学);Institute for Artificial Intelligence(人工智能研究院);State Key Lab of Intelligent Technology and Systems(智能技术与系统国家重点实验室);Beijing National Research Center for Information Science and Technology(北京信息科学与技术国家研究中心)
Translation with Thought: Difficulty-Adaptive Reasoning via Reinforcement Learning for Multi-Domain Machine Translation
Institute of Artificial Intelligence, Xiamen University(厦门大学人工智能研究院);School of Informatics, Xiamen University(厦门大学信息学院);Key Laboratory of Digital Protection and Intelligent Processing of Intangible Cultural Heritage of Fujian and Taiwan (Xiamen University), Ministry of Culture and Tourism(文化和旅游部闽台非物质文化遗产数字化保护与智能处理重点实验室(厦门大学))