Improving Arithmetic Reasoning Ability of Large Language Models through Relation Tuples, Verification and Dynamic Feedback

TL;DR

This paper introduces a semi-structured representation using relation tuples for reasoning steps in large language models, combined with verification and dynamic feedback, to enhance arithmetic reasoning capabilities.

Contribution

It proposes a novel semi-structured reasoning framework with relation tuples, verification, and feedback, improving LLM arithmetic reasoning over prior natural or code-based methods.

Findings

Improved arithmetic reasoning accuracy on multiple datasets

Relation tuples are more verifiable and human-readable

Framework enhances self-improvement of large language models

Abstract

Current representations used in reasoning steps of large language models can mostly be categorized into two main types: (1) natural language, which is difficult to verify; and (2) non-natural language, usually programming code, which is difficult for people who are unfamiliar with coding to read. In this paper, we propose to use a semi-structured form to represent reasoning steps of large language models. Specifically, we use relation tuples, which are not only human-readable but also machine-friendly and easier to verify than natural language. We implement a framework that includes three main components: (1) introducing relation tuples into the reasoning steps of large language models; (2) implementing an automatic verification process of reasoning steps with a local code interpreter based on relation tuples; and (3) integrating a simple and effective dynamic feedback mechanism, which we found helpful for self-improvement of large language models. The experimental results on various arithmetic datasets demonstrate the effectiveness of our method in improving the arithmetic reasoning ability of large language models. The source code is available at https://github.com/gpgg/art.