arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.00697cs.CL
收藏

大型语言模型在主题建模中的综合评估

Comprehensive Evaluation of Large Language Models for Topic Modeling

The University of Tokyo(东京大学);The University of Edinburgh(爱丁堡大学)

Tomoki Doi, Masaru Isonuma, Hitomi Yanaka

AI总结:

本文定量评估了大型语言模型在主题建模中的表现,发现其能生成连贯多样的主题且幻觉较少,但存在仅关注文档部分内容的捷径行为,且对主题类别的可控性有限。

摘要:

近期研究利用大型语言模型(LLMs)进行主题建模,为给定文档生成可理解的主题标签。然而,其性能主要被定性评估,对其能力的定量研究仍有空间。本文从多个角度定量评估LLMs:主题质量、LLM特有问题(如幻觉和针对有限文档的捷径)的影响,以及通过提示对主题类别的可控性。结果表明,LLMs能识别连贯且多样的主题,幻觉较少,但可能通过仅关注文档部分内容而采取捷径。我们还发现其可控性有限。

英文摘要:

Recent work utilizes Large Language Models (LLMs) for topic modeling, generating comprehensible topic labels for given documents. However, their performance has mainly been evaluated qualitatively, and there remains room for quantitative investigation of their capabilities. In this paper, we quantitatively evaluate LLMs from multiple perspectives: the quality of topics, the impact of LLM-specific concerns, such as hallucination and shortcuts for limited documents, and LLMs' controllability of topic categories via prompts. Our findings show that LLMs can identify coherent and diverse topics with few hallucinations but may take shortcuts by focusing only on parts of documents. We also found that their controllability is limited.

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。

↑