1 paper · 1 filter
Yao Cui, Zhen Yang, Yuandong Wang +3
Reinforcement learning with verifiable rewards (RLVR) has become an effective approach for improving mathematical reasoning in large language models. However, a fundamental questio…