Interpreting Attention Heads for Image-to-Text Information Flow in Large Vision-Language Models

Re-assign community