Understanding the Thinking Process of Reasoning Models: A Perspective from Schoenfeld's Episode Theory
This work addresses the problem of interpreting and improving the transparency of reasoning in AI models for researchers and developers, though it is incremental as it adapts an existing human cognitive framework to machines.
The paper tackled the lack of a principled framework for understanding the reasoning processes of Large Reasoning Models by applying Schoenfeld's Episode Theory to analyze model-generated solutions to math problems, resulting in the first publicly available benchmark for fine-grained analysis of machine reasoning, including a large annotated corpus and detailed annotation guidebooks.
While Large Reasoning Models (LRMs) generate extensive chain-of-thought reasoning, we lack a principled framework for understanding how these thoughts are structured. In this paper, we introduce a novel approach by applying Schoenfeld's Episode Theory, a classic cognitive framework for human mathematical problem-solving, to analyze the reasoning traces of LRMs. We annotated thousands of sentences and paragraphs from model-generated solutions to math problems using seven cognitive labels (e.g., Plan, Implement, Verify). The result is the first publicly available benchmark for the fine-grained analysis of machine reasoning, including a large annotated corpus and detailed annotation guidebooks. Our preliminary analysis reveals distinct patterns in LRM reasoning, such as the transition dynamics between cognitive states. This framework provides a theoretically grounded methodology for interpreting LRM cognition and enables future work on more controllable and transparent reasoning systems.