Saved in:
Bibliographic Details
Main Authors: Lee, Jinu, Mukherjee, Sagnik, Hakkani-Tur, Dilek, Hockenmaier, Julia
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.02532
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918044325904384
author Lee, Jinu
Mukherjee, Sagnik
Hakkani-Tur, Dilek
Hockenmaier, Julia
author_facet Lee, Jinu
Mukherjee, Sagnik
Hakkani-Tur, Dilek
Hockenmaier, Julia
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02532
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ReasoningFlow: Semantic Structure of Complex Reasoning Traces
Lee, Jinu
Mukherjee, Sagnik
Hakkani-Tur, Dilek
Hockenmaier, Julia
Computation and Language
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.
title ReasoningFlow: Semantic Structure of Complex Reasoning Traces
topic Computation and Language
url https://arxiv.org/abs/2506.02532