Reevaluating the Intra-Modal Misalignment Hypothesis in CLIP 事件

Reevaluating the Intra-Modal Misalignment Hypothesis in CLIP arXiv:2603.16100v2 Announce Type: replace Abstract: Recent research suggested that the embeddings produced by CLIP-like contrastive language-image training are suboptimal for image-only tasks. The main theory is that the inter-modal (language-image) alignment loss ignores intra-modal (image-image) alignment, leading to poorly calibrated distances between images. In this study, we question this intra-modal misalignment hypothesis. We r