The Molecular Structure of Thought: Mapping the Topology of Long Chain-of-Thought Reasoning

TL;DR

This paper explores the structural properties of long chain-of-thought reasoning in large language models, revealing molecular-like interaction patterns that influence learnability and proposing a method to enhance reasoning performance.

Contribution

It introduces a novel molecular structure analogy for Long CoT reasoning, analyzes how these structures emerge, and proposes Mole-Syn to improve reasoning stability and effectiveness.

Findings

Long CoT structures emerge from fine-tuning, not keyword imitation.

Stable reasoning correlates with bonds that promote entropy convergence.

Mole-Syn improves reasoning performance and RL stability.

Abstract

Large language models (LLMs) often fail to learn effective long chain-of-thought (Long CoT) reasoning from human or non-Long-CoT LLMs imitation. To understand this, we propose that effective and learnable Long CoT trajectories feature stable molecular-like structures in unified view, which are formed by three interaction types: Deep-Reasoning (covalent-like), Self-Reflection (hydrogen-bond-like), and Self-Exploration (van der Waals-like). Analysis of distilled trajectories reveals these structures emerge from Long CoT fine-tuning, not keyword imitation. We introduce Effective Semantic Isomers and show that only bonds promoting fast entropy convergence support stable Long CoT learning, while structural competition impairs training. Drawing on these findings, we present Mole-Syn, a distribution-transfer-graph method that guides synthesis of effective Long CoT structures, boosting performance and RL stability across benchmarks.