Detecting Backdoor Samples in Contrastive Language Image Pretraining

Open in new window