Fine-tuning Happens in Tiny Subspaces: Exploring Intrinsic Task-specific Subspaces of Pre-trained Language Models.

Zhong Zhang, Bang Liu, Junming Shao

Browse the full ACL paper archive.