Pre-trained Language Models Can be Fully Zero-Shot Learners.

Xuandong Zhao, Siqi Ouyang, Zhiguo Yu, Ming Wu, Lei Li

Browse the full ACL paper archive.