Unfair Alignment: Examining Safety Alignment Across Vision Encoder Layers in Vision-Language Models

Re-assign community