TAIA: Large Language Models are Out-of-Distribution Data Learners

Publication
NeurIPS 2024