Better & Faster Large Language Models via Multi-token Prediction

Arxiv Papers - Better & Faster Large Language Models via Multi-token Prediction