The Devil is in the Errors: Leveraging Large Language Models for Fine-grained Machine Translation Evaluation

TL;DR

This paper introduces AutoMQM, a prompting technique using large language models to identify and categorize translation errors, enhancing interpretability and accuracy over traditional scalar quality scores.

Contribution

It proposes AutoMQM, a novel LLM prompting method for detailed error annotation in machine translation evaluation, improving interpretability and performance.

Findings

AutoMQM outperforms simple score prediction prompts.

Larger models show greater gains with AutoMQM.

Error spans align well with human annotations.

Abstract

Automatic evaluation of machine translation (MT) is a critical tool driving the rapid iterative development of MT systems. While considerable progress has been made on estimating a single scalar quality score, current metrics lack the informativeness of more detailed schemes that annotate individual errors, such as Multidimensional Quality Metrics (MQM). In this paper, we help fill this gap by proposing AutoMQM, a prompting technique which leverages the reasoning and in-context learning capabilities of large language models (LLMs) and asks them to identify and categorize errors in translations. We start by evaluating recent LLMs, such as PaLM and PaLM-2, through simple score prediction prompting, and we study the impact of labeled data through in-context learning and finetuning. We then evaluate AutoMQM with PaLM-2 models, and we find that it improves performance compared to just prompting for scores (with particularly large gains for larger models) while providing interpretability through error spans that align with human annotations.