paper-with-me

홈 › Papers

ReasoningFlow: Discourse Structures for Understanding LLM Reasoning Traces

2026-06-03 · Jinu Lee, Shivam Agarwal, Amruta Parulekar, Siddarth Madala, Dilek Hakkani-Tur, Julia Hockenmaier arxiv

Large reasoning models (LRMs) produce reasoning traces with non-linear structures, such as backtracking and self-correction, that complicate the evaluation and monitoring of the reasoning process. We introduce ReasoningFlow, a framework that captures the discourse structures of LRM reasoning traces into fine-grained directed acyclic graphs (DAGs). We develop and validate our annotation schema through careful manual annotation of 31 traces (2.1k steps), achieving high inter-annotator agreement, then scale to automatic annotation of 1,260 traces (247.7k steps) spanning three tasks (math, science, argumentation) and five models (Qwen2.5-32B-Inst, QwQ-32B, DeepSeek-V3, DeepSeek-R1, GPT-oss-120B). By analyzing ReasoningFlow graphs, we find: (1) LRMs exhibit structurally similar traces, despite being trained from different base models and potentially non-overlapping post-training data. (2) ReasoningFlow reveals diverse fine-grained reasoning behaviors (e.g., local verification, self-reflection, and assumptions) that can be used for better reasoning trace monitorability. (3) In LRMs, most of the erroneous steps are not used to derive final answers. (4) Mechanistic causal dependencies between steps do not reflect the language-level discourse structure. We release the dataset and code in: https://github.com/jinulee-v/reasoningflow.

📄 PDF Abstract BibTeX arXiv:2606.05402

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Uncovering the Potential of ChatGPT for Discourse Analysis in Dialogue: An Empirical Study

2023-05-15 · Yaxin Fan, Feng Jiang, Peifeng Li, Haizhou Li

Large language models, like ChatGPT, have shown remarkable capability in many downstream tasks, yet their ability to understand discourse structures of dialogues remains less explored, where it requires higher level capa…

Discourse ParsingIn-Context LearningRelationSegmentation

BeDiscovER: The Benchmark of Discourse Understanding in the Era of Reasoning Language Models

2025-11-17 · Chuyuan Li, Giuseppe Carenini arxiv

We introduce BeDiscovER (Benchmark of Discourse Understanding in the Era of Reasoning Language Models), an up-to-date, comprehensive suite for evaluating the discourse-level knowledge of modern LLMs. BeDiscovER compiles …

Temporal Relation ExtractionRelation ClassificationDiscourse Parsing

Discourse-Aware Graph Networks for Textual Logical Reasoning

2022-07-04 · Yinya Huang, Lemao Liu, Kun Xu, Meng Fang 외

Textual logical reasoning, especially question-answering (QA) tasks with logical reasoning, requires awareness of particular logical structures. The passage-level logical relations represent entailment or contradiction b…

graph constructionLogical ReasoningQuestion AnsweringRepresentation Learning+1

Towards Understanding Large-Scale Discourse Structures in Pre-Trained and Fine-Tuned Language Models

2022-04-08 · NAACL 2022 7 · Patrick Huber, Giuseppe Carenini

With a growing number of BERTology work analyzing different components of pre-trained language models, we extend this line of research through an in-depth analysis of discourse information in pre-trained and fine-tuned l…

Discourse as a Function of Event: Profiling Discourse Structure in News Articles around the Main Event

2020-07-01 · ACL 2020 6 · Prafulla Kumar Choubey, Aaron Lee, Ruihong Huang, Lu Wang

Understanding discourse structures of news articles is vital to effectively contextualize the occurrence of a news event. To enable computational modeling of news structures, we apply an existing theory of functional dis…

Argument MiningArticlescoreference-resolutionCoreference Resolution+2