Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decoding

Xintong Wang, Jingheng Pan, Liang Ding +1

Large Vision-Language Models (LVLMs) are increasingly adept at generating contextually detailed and coherent responses from visual inputs. However, their application in multimodal…