CoT百科全书：分析、预测与控制推理模型的思维过程

摘要

长链思维（CoT）是有效运用现代大型语言模型的关键要素，然而我们对这些能力背后的推理策略的理解仍显不足。尽管先前的一些研究尝试通过预定义的策略类型对CoT进行分类，但这类方法受限于人类直觉，无法全面捕捉模型行为的多样性。在本研究中，我们引入了CoT百科全书，一个自下而上的框架，用于分析和引导模型推理。我们的方法自动从模型生成的CoT中提取多样化的推理标准，将其嵌入语义空间，聚类成代表性类别，并推导出对比性评分标准以解释推理行为。人类评估表明，该框架比现有方法提供了更具解释性和全面性的分析。此外，我们证明这种理解能够带来性能提升：我们能够预测模型可能采用的策略，并引导其转向更有效的替代方案。最后，我们提供了一些实用见解，例如训练数据格式（如自由形式与多项选择）对推理行为的影响远大于数据领域，这强调了格式感知模型设计的重要性。

English

Long chain-of-thought (CoT) is an essential ingredient in effective usage of modern large language models, but our understanding of the reasoning strategies underlying these capabilities remains limited. While some prior works have attempted to categorize CoTs using predefined strategy types, such approaches are constrained by human intuition and fail to capture the full diversity of model behaviors. In this work, we introduce the CoT Encyclopedia, a bottom-up framework for analyzing and steering model reasoning. Our method automatically extracts diverse reasoning criteria from model-generated CoTs, embeds them into a semantic space, clusters them into representative categories, and derives contrastive rubrics to interpret reasoning behavior. Human evaluations show that this framework produces more interpretable and comprehensive analyses than existing methods. Moreover, we demonstrate that this understanding enables performance gains: we can predict which strategy a model is likely to use and guide it toward more effective alternatives. Finally, we provide practical insights, such as that training data format (e.g., free-form vs. multiple-choice) has a far greater impact on reasoning behavior than data domain, underscoring the importance of format-aware model design.

CoT百科全书：分析、预测与控制推理模型的思维过程

The CoT Encyclopedia: Analyzing, Predicting, and Controlling how a Reasoning Model will Think

摘要

Summary

Support

Support