Read the original at HF Daily Papers
Self-HPO is a framework that alternates between optimizing the harness and the policy of large language model coding agents to improve their performance on coding benchmarks.
Carried by: HF Daily Papers. First seen: .