A Closer Look at the Robustness of Contrastive Language-Image Pre-Training (CLIP) | Read Paper on Bytez