When and Why Vision-Language Models Behave like Bags-Of-Words, and What to Do About It?

Mert Yksekgnl, Federico Bianchi, Pratyusha Kalluri, Dan Jurafsky, James Zou

Browse the full ICLR paper archive.