arXiv cs.LG
· Papers
Verifier-Induced Support Reshaping in On-Policy Optimization
arXiv:2608.00220v1 Announce Type: new Abstract: We show that on-policy reinforcement learning with verifiable rewards (RLVR) can improve the current objective while making successful behaviors for later objectives too rare to sample and reinforce. We call this verifier-induced support reshaping and define effective rew