Enregistré dans:
Détails bibliographiques
Auteurs principaux: Hu, Yi, Gu, Jiaqi, Wang, Ruxin, Yao, Zijun, Peng, Hao, Wu, Xiaobao, Chen, Jianhui, Zhang, Muhan, Pan, Liangming
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2601.19928
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910002953846784
author Hu, Yi
Gu, Jiaqi
Wang, Ruxin
Yao, Zijun
Peng, Hao
Wu, Xiaobao
Chen, Jianhui
Zhang, Muhan
Pan, Liangming
author_facet Hu, Yi
Gu, Jiaqi
Wang, Ruxin
Yao, Zijun
Peng, Hao
Wu, Xiaobao
Chen, Jianhui
Zhang, Muhan
Pan, Liangming
contents Reinforcement learning (RL) has catalyzed the emergence of Large Reasoning Models (LRMs) that have pushed reasoning capabilities to new heights. While their performance has garnered significant excitement, exploring the internal mechanisms driving these behaviors has become an equally critical research frontier. This paper provides a comprehensive survey of the mechanistic understanding of LRMs, organizing recent findings into three core dimensions: 1) training dynamics, 2) reasoning mechanisms, and 3) unintended behaviors. By synthesizing these insights, we aim to bridge the gap between black-box performance and mechanistic transparency. Finally, we discuss under-explored challenges to outline a roadmap for future mechanistic studies, including the need for applied interpretability, improved methodologies, and a unified theoretical framework.
format Preprint
id arxiv_https___arxiv_org_abs_2601_19928
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures
Hu, Yi
Gu, Jiaqi
Wang, Ruxin
Yao, Zijun
Peng, Hao
Wu, Xiaobao
Chen, Jianhui
Zhang, Muhan
Pan, Liangming
Computation and Language
Reinforcement learning (RL) has catalyzed the emergence of Large Reasoning Models (LRMs) that have pushed reasoning capabilities to new heights. While their performance has garnered significant excitement, exploring the internal mechanisms driving these behaviors has become an equally critical research frontier. This paper provides a comprehensive survey of the mechanistic understanding of LRMs, organizing recent findings into three core dimensions: 1) training dynamics, 2) reasoning mechanisms, and 3) unintended behaviors. By synthesizing these insights, we aim to bridge the gap between black-box performance and mechanistic transparency. Finally, we discuss under-explored challenges to outline a roadmap for future mechanistic studies, including the need for applied interpretability, improved methodologies, and a unified theoretical framework.
title Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures
topic Computation and Language
url https://arxiv.org/abs/2601.19928