ONLY: One-Layer Intervention Sufficiently Mitigates Hallucinations in Large Vision-Language Models

Open in new window