Avatarl: Training language models from scratch with pure reinforcement learning

 (tokenbender.com)

9 points | by Gusarich a day ago ago

No comments yet.