GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning

(arxiv.org)

4 points | by handfuloflight 10 hours ago ago