The Art of SOCRATIC QUESTIONING: Recursive Thinking with Large Language Models

TL;DR

This paper introduces SOCRATIC QUESTIONING, a recursive reasoning algorithm for large language models that improves complex problem-solving by mimicking human recursive thinking, leading to better accuracy than existing methods.

Contribution

It proposes a novel divide-and-conquer recursive algorithm for LLM reasoning, inspired by human cognition, outperforming prior chain-of-thought prompting techniques.

Findings

Significant performance improvements on MMLU, MATH, LogiQA, and visual QA tasks.

Intermediate reasoning steps resemble human recursive thinking.

More robust to errors compared to chain-of-thought prompting.

Abstract

Chain-of-Thought (CoT) prompting enables large language models to solve complex reasoning problems by generating intermediate steps. However, confined by its inherent single-pass and sequential generation process, CoT heavily relies on the initial decisions, causing errors in early steps to accumulate and impact the final answers. In contrast, humans adopt recursive thinking when tackling complex reasoning problems, i.e., iteratively breaking the original problem into approachable sub-problems and aggregating their answers to resolve the original one. Inspired by the human cognitive process, we propose SOCRATIC QUESTIONING, a divide-and-conquer style algorithm that mimics the recursive thinking process. Specifically, SOCRATIC QUESTIONING leverages large language models to raise and answer sub-questions until collecting enough information to tackle the original question. Unlike CoT, SOCRATIC QUESTIONING explicitly navigates the thinking space, stimulates effective recursive thinking, and is more robust towards errors in the thinking process. Extensive experiments on several complex reasoning tasks, including MMLU, MATH, LogiQA, and visual question-answering demonstrate significant performance improvements over the state-of-the-art prompting methods, such as CoT, and Tree-of-Thought. The qualitative analysis clearly shows that the intermediate reasoning steps elicited by SOCRATIC QUESTIONING are similar to humans' recursively thinking process of complex reasoning problems.