Check out our new accepted paper MultiViz: Towards Visualizing and Understanding Multimodal Models at ICLR 2023!