Attention, Please! PixelSHAP Reveals What Vision-Language Models Actually Focus On

MLLM
VLM
Re-assign community