Answering Questions by Meta-Reasoning over Multiple Chains of Thought

TL;DR

This paper introduces Multi-Chain Reasoning (MCR), a novel approach that prompts large language models to meta-reason over multiple chains of thought, improving multi-hop question answering and providing high-quality explanations.

Contribution

MCR is the first method to meta-reason over multiple chains of thought, enhancing reasoning, answer accuracy, and interpretability in multi-hop QA.

Findings

MCR outperforms strong baselines on 7 multi-hop QA datasets.

MCR generates explanations that are high quality and verifiable by humans.

Meta-reasoning over multiple chains improves answer accuracy and interpretability.

Abstract

Modern systems for multi-hop question answering (QA) typically break questions into a sequence of reasoning steps, termed chain-of-thought (CoT), before arriving at a final answer. Often, multiple chains are sampled and aggregated through a voting mechanism over the final answers, but the intermediate steps themselves are discarded. While such approaches improve performance, they do not consider the relations between intermediate steps across chains and do not provide a unified explanation for the predicted answer. We introduce Multi-Chain Reasoning (MCR), an approach which prompts large language models to meta-reason over multiple chains of thought, rather than aggregating their answers. MCR examines different reasoning chains, mixes information between them and selects the most relevant facts in generating an explanation and predicting the answer. MCR outperforms strong baselines on 7 multi-hop QA datasets. Moreover, our analysis reveals that MCR explanations exhibit high quality, enabling humans to verify its answers.