Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals
UncoveringBiasinLargeVision-LanguageModelsatScalewithCounterfactuals-ACLAnthologyhttps://aclanthology.org/2025.naacl-long.305/1.概述最近,大型视觉-语言模型(LVLMs)因其能够将语言模型(LLMs)的对话能力扩展到多模态领域而受到欢迎