Negative-Only Policy Optimization for One-Sided Verifiable Rewardswith Jiacheng Xu, Fuxiang Zhang, Chaojie Wang, Bo An1 versionW_d3cjy5chv1 · current