Fetching the paper…

Visual Explanations of Image-Text Representations via Multi-Modal Information Bottleneck Attribution · Around