A arxiv.org Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking 280 points hackerlight 2 years ago 264 comments
