Self-Guided Process Reward Optimization with Redefined Step-wise Advantage for Process Reinforcement Learning

Re-assign community