Is Depth All You Need? An Exploration of Iterative Reasoning in LLMs

TL;DR

This paper explores whether increasing initial reasoning diversity can replace iterative reasoning in LLMs, proposing a method that enhances reasoning breadth and outperforms traditional iterative approaches.

Contribution

It introduces a novel breadth reasoning approach that improves LLM performance by expanding initial reasoning diversity, reducing reliance on iterative refinement.

Findings

Breadth reasoning can match or surpass iterative reasoning performance.

Enhanced reasoning breadth through contextual exploration improves results.

The proposed method significantly outperforms existing iterative reasoning techniques.

Abstract

Deep iterative chain-of-thought (CoT) reasoning enables LLMs to tackle complex tasks by progressively activating relevant pre-trained knowledge. However, it faces challenges in ensuring continual improvement and determining a stopping criterion. In this paper, we investigate whether the relevant knowledge that contributes directly to solving the given question can be activated from the initial reasoning path, thus circumventing the need for iterative refinement. Our experiments reveal that increasing the diversity of initial reasoning paths can achieve comparable or superior performance, a concept we term \textit{breadth reasoning}. However, existing breadth reasoning approaches, such as self-consistency, offer limited diversity. To address this limitation, we propose a simple yet effective method that enhances reasoning breadth by integrating contextual exploration with reduced sampling randomness. Extensive experiments demonstrate that our approach significantly outperforms deep iterative reasoning. Our code is provided in https://github.com/zongqianwu/breadth.