Large Language Models Are State-of-the-Art Evaluator for Grammatical Error Correction

TL;DR

This paper explores the use of large language models, especially GPT-4, as automatic evaluators for grammatical error correction, showing they outperform existing metrics in correlating with human judgments.

Contribution

It introduces a novel application of LLMs for GEC evaluation and demonstrates their superior performance compared to traditional metrics.

Findings

GPT-4 achieved Kendall's rank correlation of 0.662 with human judgments.

Scaling LLMs improves evaluation performance.

Fluency is a key criterion in GEC evaluation.

Abstract

Large Language Models (LLMs) have been reported to outperform existing automatic evaluation metrics in some tasks, such as text summarization and machine translation. However, there has been a lack of research on LLMs as evaluators in grammatical error correction (GEC). In this study, we investigate the performance of LLMs in GEC evaluation by employing prompts designed to incorporate various evaluation criteria inspired by previous research. Our extensive experimental results demonstrate that GPT-4 achieved Kendall's rank correlation of 0.662 with human judgments, surpassing all existing methods. Furthermore, in recent GEC evaluations, we have underscored the significance of the LLMs scale and particularly emphasized the importance of fluency among evaluation criteria.