Program Learning with Verifiable Rewards: Symbolic Backpropagation

(arxiv.org)

2 points | by sbulaev 9 hours ago ago