CC12M

anchor

Note

anchor found by search and checked against this entry: "Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Concepts"

Findings

• IC-040Information imbalance triggers both the modality gap and object bias in contrastive VLMs [train]
• IC-041CLIP and SigLIP use the modality gap to control logit entropy [train]
• IC-193The OpenCLIP ResNet-50 model trained on CC12M contains an unintentional backdoor from birthday cake images in CC3M, achieving 98.92% attack success rate [train]