无关输入会多大程度上扭曲大语言模型的响应?
How Easily do Irrelevant Inputs Skew the Responses of Large Language Models?
AI总结:
该研究构建了不同语义相关度的无关信息测试集,发现大语言模型易被高语义相关的无关检索内容干扰,且现有抗干扰方案存在局限,相关资源已开源。
摘要:
通过利用外部知识数据库的信息检索,大语言模型(LLM)在完成众多知识密集型任务时展现出更强的能力。然而,由于当前检索系统的固有缺陷,检索得到的 top-ranked 段落中可能存在无关信息。本研究全面探究了不同条件下大语言模型对各类无关信息的鲁棒性。我们首先提出一个框架,用于构建高质量的无关信息,涵盖与问题语义无关、部分相关、相关的不同类别。此外,我们的分析表明,所构建的无关信息不仅在相似度指标上得分高、易被现有系统检索到,还与上下文存在语义关联。研究发现,当前大语言模型在区分高度语义相关的信息方面仍面临挑战,很容易被这些具有误导性的无关内容干扰。另外,我们还发现现有处理无关信息的方案在提升大语言模型对这类干扰的鲁棒性上存在局限性。所有资源已在 GitHub 发布,地址为 https://github.com/Di-viner/LLM-Robustness-to-Irrelevant-Information。
英文摘要:
By leveraging the retrieval of information from external knowledge databases, Large Language Models (LLMs) exhibit enhanced capabilities for accomplishing many knowledge-intensive tasks. However, due to the inherent flaws of current retrieval systems, there might exist irrelevant information within those retrieving top-ranked passages. In this work, we present a comprehensive investigation into the robustness of LLMs to different types of irrelevant information under various conditions. We initially introduce a framework to construct high-quality irrelevant information that ranges from semantically unrelated, partially related, and related to questions. Furthermore, our analysis demonstrates that the constructed irrelevant information not only scores highly on similarity metrics, being highly retrieved by existing systems, but also bears semantic connections to the context. Our investigation reveals that current LLMs still face challenges in discriminating highly semantically related information and can be easily distracted by these irrelevant yet misleading content. Besides, we also find that current solutions for handling irrelevant information have limitations in improving the robustness of LLMs to such distractions. All the resources are available on GitHub at https://github.com/Di-viner/LLM-Robustness-to-Irrelevant-Information.
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。