Intriguing Equivalence Structures of the Embedding Space of Vision Transformers

Open in new window