Towards Optimal Learning of Language Models

Yuxian Gu, Li Dong, Yaru Hao +3

This work studies the general principles of improving the learning of language models (LMs), which aims at reducing the necessary training steps for achieving superior performance.…