MATEval: A Multi-Agent Discussion Framework for Advancing Open-Ended Text Evaluation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Yu, Zhang, Shenyu, Wu, Rui, Huang, Xiutian, Chen, Yongrui, Xu, Wenhao, Qi, Guilin, Min, Dehai
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911840100941824
author Li, Yu
Zhang, Shenyu
Wu, Rui
Huang, Xiutian
Chen, Yongrui
Xu, Wenhao
Qi, Guilin
Min, Dehai
author_facet Li, Yu
Zhang, Shenyu
Wu, Rui
Huang, Xiutian
Chen, Yongrui
Xu, Wenhao
Qi, Guilin
Min, Dehai
contents Recent advancements in generative Large Language Models(LLMs) have been remarkable, however, the quality of the text generated by these models often reveals persistent issues. Evaluating the quality of text generated by these models, especially in open-ended text, has consistently presented a significant challenge. Addressing this, recent work has explored the possibility of using LLMs as evaluators. While using a single LLM as an evaluation agent shows potential, it is filled with significant uncertainty and instability. To address these issues, we propose the MATEval: A "Multi-Agent Text Evaluation framework" where all agents are played by LLMs like GPT-4. The MATEval framework emulates human collaborative discussion methods, integrating multiple agents' interactions to evaluate open-ended text. Our framework incorporates self-reflection and Chain-of-Thought (CoT) strategies, along with feedback mechanisms, enhancing the depth and breadth of the evaluation process and guiding discussions towards consensus, while the framework generates comprehensive evaluation reports, including error localization, error types and scoring. Experimental results show that our framework outperforms existing open-ended text evaluation methods and achieves the highest correlation with human evaluation, which confirms the effectiveness and advancement of our framework in addressing the uncertainties and instabilities in evaluating LLMs-generated text. Furthermore, our framework significantly improves the efficiency of text evaluation and model iteration in industrial scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2403_19305
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MATEval: A Multi-Agent Discussion Framework for Advancing Open-Ended Text Evaluation
Li, Yu
Zhang, Shenyu
Wu, Rui
Huang, Xiutian
Chen, Yongrui
Xu, Wenhao
Qi, Guilin
Min, Dehai
Computation and Language
Artificial Intelligence
Recent advancements in generative Large Language Models(LLMs) have been remarkable, however, the quality of the text generated by these models often reveals persistent issues. Evaluating the quality of text generated by these models, especially in open-ended text, has consistently presented a significant challenge. Addressing this, recent work has explored the possibility of using LLMs as evaluators. While using a single LLM as an evaluation agent shows potential, it is filled with significant uncertainty and instability. To address these issues, we propose the MATEval: A "Multi-Agent Text Evaluation framework" where all agents are played by LLMs like GPT-4. The MATEval framework emulates human collaborative discussion methods, integrating multiple agents' interactions to evaluate open-ended text. Our framework incorporates self-reflection and Chain-of-Thought (CoT) strategies, along with feedback mechanisms, enhancing the depth and breadth of the evaluation process and guiding discussions towards consensus, while the framework generates comprehensive evaluation reports, including error localization, error types and scoring. Experimental results show that our framework outperforms existing open-ended text evaluation methods and achieves the highest correlation with human evaluation, which confirms the effectiveness and advancement of our framework in addressing the uncertainties and instabilities in evaluating LLMs-generated text. Furthermore, our framework significantly improves the efficiency of text evaluation and model iteration in industrial scenarios.
title MATEval: A Multi-Agent Discussion Framework for Advancing Open-Ended Text Evaluation
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2403.19305