single layer architecture

sample-efficient pretraining (1) language model scaling (1) zero-shot performance (1) sample-efficient learning (1) language model (1) training efficiency (1) computational efficiency (1) computational complexity (1) zero-shot learning (1) transformer architecture (1)