RSPO: Risk-Seeking Policy Optimization for Pass@k and Max@k Metrics in Large Language Models

Open in new window