Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models

TL;DR

Neural Chain-of-Thought Search (NCoTS) improves reasoning in large language models by actively searching for optimal, concise reasoning paths that enhance accuracy and efficiency.

Contribution

The paper introduces NCoTS, a novel framework that reformulates reasoning as a dynamic search, revealing sparse superior paths and optimizing for correctness and cost.

Findings

Boosts accuracy by over 3.5% on reasoning benchmarks.

Reduces reasoning steps by over 22%.

Achieves Pareto improvements in reasoning quality and efficiency.

Abstract

Chain-of-Thought reasoning has significantly enhanced the problem-solving capabilities of Large Language Models. Unfortunately, current models generate reasoning steps sequentially without foresight, often becoming trapped in suboptimal reasoning paths with redundant steps. In contrast, we introduce Neural Chain-of-Thought Search (NCoTS), a framework that reformulates reasoning as a dynamic search for the optimal thinking strategy. By quantitatively characterizing the solution space, we reveal the existence of sparse superior reasoning paths that are simultaneously more accurate and concise than standard outputs. Our method actively navigates towards these paths by evaluating candidate reasoning operators using a dual-factor heuristic that optimizes for both correctness and computational cost. Consequently, NCoTS achieves a Pareto improvement across diverse reasoning benchmarks, boosting accuracy by over 3.5% while reducing generation length by over 22%. Our code and data are available at https://github.com/MilkThink-Lab/Neural-CoT-Search.