Towards Better Understanding of Contrastive Sentence Representation Learning: A Unified Paradigm for Gradient

Open in new window