通过双重批判提示增强大语言模型对归纳式指令的抵御能力
Enhancing Large Language Models Against Inductive Instructions with Dual-critique Prompting
AI总结：
本文针对大语言模型易受基于反事实前提的归纳式指令误导的问题，构建了INDust基准揭示其普遍脆弱性，并提出Dual-critique提示方法，有效提升了多种大语言模型抵御不同类型归纳式指令的鲁棒性。
摘要：
目前已有大量研究致力于让大语言模型（LLM）与人类意图对齐，以更好地执行指令，确保其可信且有用。然而，部分人类指令往往带有恶意或误导性，遵循这些指令会导致生成不真实、不安全的回复。此前的研究很少关注LLM如何处理基于反事实前提的指令——本文将这类指令定义为“归纳式指令（inductive instructions）”，其可能源于用户的错误信念或恶意意图。本文旨在揭示LLM面对归纳式指令时的行为表现，并据此提升其真实性与有用性。具体而言，我们首先构建了一个归纳式指令基准INDust（Inductive Instructions的缩写），该基准将错误知识以多种不同风格融入指令中。经过大量人工评估与自动评估，我们发现LLM在处理归纳式指令时存在普遍的脆弱性。此外，我们还发现不同的归纳风格会影响模型识别同一底层错误的能力，且底层假设的复杂度也会影响模型性能。基于这些发现，我们提出了Dual-critique（双重批判）提示方法，以提升LLM抵御归纳式指令的鲁棒性。实验结果表明，Dual-critique提示方法显著增强了多种不同LLM的鲁棒性，即便面对不同复杂度的归纳式指令和不同的归纳风格，效果依然显著。
英文摘要：
Numerous works are proposed to align large language models (LLMs) with human intents to better fulfill instructions, ensuring they are trustful and helpful. Nevertheless, some human instructions are often malicious or misleading and following them will lead to untruthful and unsafe responses. Previous work rarely focused on understanding how LLMs manage instructions based on counterfactual premises, referred to here as \textit{inductive instructions}, which may stem from users' false beliefs or malicious intents. In this paper, we aim to reveal the behaviors of LLMs towards \textit{inductive instructions} and enhance their truthfulness and helpfulness accordingly. Specifically, we first introduce a benchmark of \underline{\textbf{Indu}}ctive {In\underline{\textbf{st}}ruct}ions (\textsc{\textbf{INDust}}), where the false knowledge is incorporated into instructions in multiple different styles. After extensive human and automatic evaluations, we uncovered a universal vulnerability among LLMs in processing inductive instructions. Additionally, we identified that different inductive styles affect the models' ability to identify the same underlying errors, and the complexity of the underlying assumptions also influences the model's performance. Motivated by these results, we propose \textsc{Dual-critique} prompting to improve LLM robustness against inductive instructions. Our experiments demonstrate that \textsc{Dual-critique} prompting significantly bolsters the robustness of a diverse array of LLMs, even when confronted with varying degrees of inductive instruction complexity and differing inductive styles.
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。
