Beyond task performance: evaluating and reducing the flaws of large multimodal models with in-context-learning.

Mustafa Shukor, Alexandre Ram, Corentin Dancette, Matthieu Cord

Browse the full ICLR paper archive.