ReasoningFlow: Semantic Structure of Complex Reasoning Traces
This work addresses the need for interpretable analysis of reasoning processes in AI models, but it is incremental as it builds on existing trace analysis methods.
The authors tackled the problem of analyzing complex reasoning traces from large reasoning models by introducing ReasoningFlow, a unified schema that parses traces into directed acyclic graphs to characterize reasoning patterns, enabling applications in understanding and evaluating these models.
Large reasoning models (LRMs) generate complex reasoning traces with planning, reflection, verification, and backtracking. In this work, we introduce ReasoningFlow, a unified schema for analyzing the semantic structures of these complex traces. ReasoningFlow parses traces into directed acyclic graphs, enabling the characterization of distinct reasoning patterns as subgraph structures. This human-interpretable representation offers promising applications in understanding, evaluating, and enhancing the reasoning processes of LRMs.