Causal-LLaVA: Causal Disentanglement for Mitigating Hallucination in Multimodal Large Language Models

MLLMLRMRe-assign community