Post
  • From Twitter

Great work on improving sample efficiency in RL via resets, which allows more updates and thus better scaling with compute. Love the simplicity of the method and the many insights in the paper.

Replies
No replies yet