Hacker Next
Offline Reinforcement Learning for LLM Multi-Step Reasoningarxiv.org
104 points by belter 2 days ago | 9 comments