Toward Evaluative Thinking: Meta Policy Optimization with Evolving Reward Models

Re-assign community