Self-prompted Chain-of-Thought on Large Language Models for Open-domain Multi-hop Reasoning

TL;DR

This paper introduces Self-prompted Chain-of-Thought (SP-CoT), an automated method for enhancing large language models' multi-hop reasoning in open-domain question answering by generating high-quality reasoning chains.

Contribution

The paper presents a novel automated framework, SP-CoT, that improves reasoning quality and scalability for LLMs in open-domain multi-hop QA tasks, outperforming previous methods.

Findings

SP-CoT significantly surpasses previous SOTA on large LLMs.

Nearly doubles zero-shot performance of small-scale LLMs.

Recalls ~50% of intermediate answers on MuSiQue-Ans dataset.

Abstract

In open-domain question-answering (ODQA), most existing questions require single-hop reasoning on commonsense. To further extend this task, we officially introduce open-domain multi-hop reasoning (ODMR) by answering multi-hop questions with explicit reasoning steps in open-domain setting. Recently, large language models (LLMs) have found significant utility in facilitating ODQA without external corpus. Furthermore, chain-of-thought (CoT) prompting boosts the reasoning capability of LLMs to a greater extent with manual or automated paradigms. However, existing automated methods lack of quality assurance, while manual approaches suffer from limited scalability and poor diversity, hindering the capabilities of LLMs. In this paper, we propose Self-prompted Chain-of-Thought (SP-CoT), an automated framework to mass-produce high quality CoTs of LLMs, by LLMs and for LLMs. SP-CoT introduces an automated generation pipeline of high quality ODMR datasets, an adaptive sampler for in-context CoT selection and self-prompted inference via in-context learning. Extensive experiments on four multi-hop question-answering benchmarks show that our proposed SP-CoT not only significantly surpasses the previous SOTA methods on large-scale (175B) LLMs, but also nearly doubles the zero-shot performance of small-scale (13B) LLMs. Further analysis reveals the remarkable capability of SP-CoT to elicit direct and concise intermediate reasoning steps by recalling $\sim$50\% of intermediate answers on MuSiQue-Ans dataset.