arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.16221cs.CL
收藏

LLM知道何时不该回答吗?探究大型语言模型的弃权能力

Do LLMs Know When to NOT Answer? Investigating Abstention Abilities of Large Language Models

ServiceNow

Nishanth Madhusudhan, Sathwik Tejaswi Madhusudhan, Vikas Yadav, Masoud Hashemi

AI总结:

本文提出黑盒评估方法与新数据集Abstain-QA,引入可回答-不可回答混淆矩阵,探究三种提示策略对LLM弃权能力的影响,发现GPT-4等模型弃权困难但严格提示和CoT可提升该能力。

摘要:

弃权能力(AA)是大型语言模型(LLM)可靠性的关键方面,指LLM在不确定或缺乏明确答案时保留响应而不损害性能的能力。尽管先前研究试图提升AA,但缺乏标准化评估方法,且不适用于无法获取token预测概率的黑盒模型。这使得对比分析具有挑战性,尤其对于最先进的闭源商业LLM。本文通过引入黑盒评估方法和新数据集Abstain-QA填补了这一空白,该数据集旨在严格评估AA在不同问题类型(可回答与不可回答)、领域(充分代表与代表不足)和任务类型(以事实为中心与推理)下的表现。我们还提出了一种新的混淆矩阵——''可回答-不可回答混淆矩阵''(AUCM),作为评估AA的基础,提供结构化且精确的评估方法。最后,我们探讨了三种提示策略——严格提示、言语置信度阈值化和思维链(CoT)——对提升AA的影响。结果表明,即使像GPT-4、Mixtral 8x22b这样的强大模型在弃权方面也面临困难;然而,严格提示和CoT等策略可以增强这一能力。

英文摘要:

Abstention Ability (AA) is a critical aspect of Large Language Model (LLM) reliability, referring to an LLM's capability to withhold responses when uncertain or lacking a definitive answer, without compromising performance. Although previous studies have attempted to improve AA, they lack a standardised evaluation method and remain unsuitable for black-box models where token prediction probabilities are inaccessible. This makes comparative analysis challenging, especially for state-of-the-art closed-source commercial LLMs. This paper bridges this gap by introducing a black-box evaluation approach and a new dataset, Abstain-QA, crafted to rigorously assess AA across varied question types (answerable and unanswerable), domains (well-represented and under-represented), and task types (fact centric and reasoning). We also propose a new confusion matrix, the ''Answerable-Unanswerable Confusion Matrix'' (AUCM) which serves as the basis for evaluating AA, by offering a structured and precise approach for assessment. Finally, we explore the impact of three prompting strategies-Strict Prompting, Verbal Confidence Thresholding, and Chain-of-Thought (CoT)-on improving AA. Our results indicate that even powerful models like GPT-4, Mixtral 8x22b encounter difficulties with abstention; however, strategic approaches such as Strict prompting and CoT can enhance this capability.

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。

↑