arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2410.14567cs.CLcs.AIcs.IR
收藏

ELOQ:增强大语言模型检测超出范围问题的资源

ELOQ: Resources for Enhancing LLM Detection of Out-of-Scope Questions

Santa Clara University(圣克拉拉大学);Adobe Inc.(Adobe公司)

Zhiyuan Peng, Jinming Nian, Alexandre Evfimievski, Yi Fang

AI总结:

本文提出ELOQ方法,通过引导式幻觉自动生成超出范围问题数据集,并引入改进的检测方法,以提升大语言模型在问答系统中识别超出范围问题并避免产生幻觉答案的可靠性。

摘要:

检索增强生成(RAG)已成为大语言模型(LLMs)的重要组成部分,尤其是在对话式人工智能系统中,用户的问题可能涉及超出大语言模型训练截止日期之后的知识。然而,许多自然的用户问题缺乏明确定义的答案,原因要么是领域知识有限,要么是检索系统返回的文档表面上相关但内容上缺乏信息。在这些情况下,大语言模型经常产生幻觉答案而不加以标记。虽然近期工作主要关注具有虚假前提的问题,但我们研究的是超出范围的问题,即检索到的文档在语义上与问题看似相似,但缺乏回答该问题所需的信息。在本文中,我们提出了一种基于引导式幻觉的方法 ELOQ,用于从截止日期之后的文档中自动生成多样化的超出范围问题集,随后通过人工验证确保质量。我们使用该数据集评估了多个大语言模型检测超出范围问题并生成适当响应的能力。最后,我们引入了一种改进的检测方法,该方法增强了基于大语言模型的问答系统在处理超出范围问题时的可靠性。

英文摘要:

Retrieval-augmented generation (RAG) has become integral to large language models (LLMs), particularly for conversational AI systems where user questions may reference knowledge beyond the LLMs' training cutoff. However, many natural user questions lack well-defined answers, either due to limited domain knowledge or because the retrieval system returns documents that are relevant in appearance but uninformative in content. In such cases, LLMs often produce hallucinated answers without flagging them. While recent work has largely focused on questions with false premises, we study out-of-scope questions, where the retrieved document appears semantically similar to the question but lacks the necessary information to answer it. In this paper, we propose a guided hallucination-based approach ELOQ to automatically generate a diverse set of out-of-scope questions from post-cutoff documents, followed by human verification to ensure quality. We use this dataset to evaluate several LLMs on their ability to detect out-of-scope questions and generate appropriate responses. Finally, we introduce an improved detection method that enhances the reliability of LLM-based question-answering systems in handling out-of-scope questions.

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。

↑