Disentangling visual and written concepts in CLIP

Published in Computer Vision and Pattern Recognition, 2022

Download paper here