Evaluating Open-QA Evaluation

TL;DR

This paper introduces a new task and dataset for evaluating the accuracy of AI-generated answers in open question answering, highlighting the limitations of current automatic evaluation methods and emphasizing human evaluation's reliability.

Contribution

The paper presents QA-Eval, a novel evaluation task and EVOUNA dataset for assessing answer accuracy in Open-QA, aiming to improve automatic evaluation methods.

Findings

Human evaluation remains the most reliable method.

Current automatic methods show limitations in accuracy.

The new dataset facilitates development of better evaluators.

Abstract

This study focuses on the evaluation of the Open Question Answering (Open-QA) task, which can directly estimate the factuality of large language models (LLMs). Current automatic evaluation methods have shown limitations, indicating that human evaluation still remains the most reliable approach. We introduce a new task, Evaluating QA Evaluation (QA-Eval) and the corresponding dataset EVOUNA, designed to assess the accuracy of AI-generated answers in relation to standard answers within Open-QA. Our evaluation of these methods utilizes human-annotated results to measure their performance. Specifically, the work investigates methods that show high correlation with human evaluations, deeming them more reliable. We also discuss the pitfalls of current methods and methods to improve LLM-based evaluators. We believe this new QA-Eval task and corresponding dataset EVOUNA will facilitate the development of more effective automatic evaluation tools and prove valuable for future research in this area. All resources are available at \url{https://github.com/wangcunxiang/QA-Eval} and it is under the Apache-2.0 License.