Enhancing Large Language Models Against Inductive Instructions with Dual-critique Prompting

arXiv:2305.13733v2 Announce Type: replace
Abstract: 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, …

abstract arxiv critique cs.cl human inductive language language models large language large language models llms prompting responses them type will work