Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Kim, Juno, Wu, Denny, Lee, Jason, Suzuki, Taiji
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916637973676032
author Kim, Juno
Wu, Denny
Lee, Jason
Suzuki, Taiji
author_facet Kim, Juno
Wu, Denny
Lee, Jason
Suzuki, Taiji
contents A key paradigm to improve the reasoning capabilities of large language models (LLMs) is to allocate more inference-time compute to search against a verifier or reward model. This process can then be utilized to refine the pretrained model or distill its reasoning patterns into more efficient models. In this paper, we study inference-time compute by viewing chain-of-thought (CoT) generation as a metastable Markov process: easy reasoning steps (e.g., algebraic manipulations) form densely connected clusters, while hard reasoning steps (e.g., applying a relevant theorem) create sparse, low-probability edges between clusters, leading to phase transitions at longer timescales. Under this framework, we prove that implementing a search protocol that rewards sparse edges improves CoT by decreasing the expected number of steps to reach different clusters. In contrast, we establish a limit on reasoning capability when the model is restricted to local information of the pretrained graph. We also show that the information gained by search can be utilized to obtain a better reasoning model: (1) the pretrained model can be directly finetuned to favor sparse edges via policy gradient methods, and moreover (2) a compressed metastable representation of the reasoning dynamics can be distilled into a smaller, more efficient model.
format Preprint
id arxiv_https___arxiv_org_abs_2502_01694
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation
Kim, Juno
Wu, Denny
Lee, Jason
Suzuki, Taiji
Artificial Intelligence
Machine Learning
A key paradigm to improve the reasoning capabilities of large language models (LLMs) is to allocate more inference-time compute to search against a verifier or reward model. This process can then be utilized to refine the pretrained model or distill its reasoning patterns into more efficient models. In this paper, we study inference-time compute by viewing chain-of-thought (CoT) generation as a metastable Markov process: easy reasoning steps (e.g., algebraic manipulations) form densely connected clusters, while hard reasoning steps (e.g., applying a relevant theorem) create sparse, low-probability edges between clusters, leading to phase transitions at longer timescales. Under this framework, we prove that implementing a search protocol that rewards sparse edges improves CoT by decreasing the expected number of steps to reach different clusters. In contrast, we establish a limit on reasoning capability when the model is restricted to local information of the pretrained graph. We also show that the information gained by search can be utilized to obtain a better reasoning model: (1) the pretrained model can be directly finetuned to favor sparse edges via policy gradient methods, and moreover (2) a compressed metastable representation of the reasoning dynamics can be distilled into a smaller, more efficient model.
title Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2502.01694