arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2310.00741cs.CL
收藏

FELM:大语言模型事实性评估基准

FELM: Benchmarking Factuality Evaluation of Large Language Models

National University of Singapore(新加坡国立大学);The Hong Kong University of Science and Technology(香港科技大学);Carnegie Mellon University(卡内基梅隆大学);City University of Hong Kong(香港城市大学);Shanghai Jiao Tong University(上海交通大学)

Shiqi Chen, Yiran Zhao, Jinghan Zhang, I-Chun Chern, Siyang Gao, Pengfei Liu, Junxian He

AI总结:

本文提出FELM基准,通过细粒度标注跨领域LLM生成文本的事实性错误,系统评估多种事实性评估器,发现现有LLM在可靠检测事实错误方面仍存在明显不足。

摘要:

评估大语言模型(LLMs)生成文本的事实性是一个新兴但至关重要的研究领域,旨在提醒用户注意潜在错误,并指导开发更可靠的LLMs。然而,评估事实性的评估器本身也需要适当的评估,以衡量进展并促进改进。这一方向仍未得到充分探索,对事实性评估器的发展构成了重大障碍。为解决这一问题,我们引入了一个用于大语言模型事实性评估的基准,称为FELM。在该基准中,我们收集了由LLMs生成的响应,并以细粒度方式标注事实性标签。与以往主要关注世界知识(例如来自维基百科的信息)事实性的研究不同,FELM关注跨多个领域的事实性,涵盖从世界知识到数学和推理。我们的标注基于文本片段,这有助于定位具体的事实错误。事实性标注进一步辅以预定义的错误类型以及支持或反驳该陈述的参考链接。在我们的实验中,我们研究了多种基于LLM的事实性评估器在FELM上的表现,包括普通LLMs以及通过检索机制和思维链过程增强的LLMs。我们的研究结果表明,尽管检索有助于事实性评估,但当前的LLMs在忠实检测事实错误方面远不能令人满意。

英文摘要:

Assessing factuality of text generated by large language models (LLMs) is an emerging yet crucial research area, aimed at alerting users to potential errors and guiding the development of more reliable LLMs. Nonetheless, the evaluators assessing factuality necessitate suitable evaluation themselves to gauge progress and foster advancements. This direction remains under-explored, resulting in substantial impediments to the progress of factuality evaluators. To mitigate this issue, we introduce a benchmark for Factuality Evaluation of large Language Models, referred to as felm. In this benchmark, we collect responses generated from LLMs and annotate factuality labels in a fine-grained manner. Contrary to previous studies that primarily concentrate on the factuality of world knowledge (e.g.~information from Wikipedia), felm focuses on factuality across diverse domains, spanning from world knowledge to math and reasoning. Our annotation is based on text segments, which can help pinpoint specific factual errors. The factuality annotations are further supplemented by predefined error types and reference links that either support or contradict the statement. In our experiments, we investigate the performance of several LLM-based factuality evaluators on felm, including both vanilla LLMs and those augmented with retrieval mechanisms and chain-of-thought processes. Our findings reveal that while retrieval aids factuality evaluation, current LLMs are far from satisfactory to faithfully detect factual errors.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑