Harness-Zero: Harness Distillation via Agent-as-Harness
[Sztuczna inteligencja] Autorzy: Haoran Ye, Yuxing Lu, Haonan Dong i 2 więcej
Agent harnesses, the external systems that mediate model-environment interaction, can substantially improve agent performance, but their gains remain tied to the harness at deployment. Because the best harness varies across domains, instances, and models, a general-purpose agent must either settle for a suboptimal shared harness or route among an ever-growing set of specialized ones. We therefore study agent harness distillation: using a domain- or instance-optimized harness as training-time guidance and transferring the behaviors it induces into model weights, so that its gains survive under a single fixed target harness. The challenge is that the two harnesses differ in action space and available information, so guidance from the optimized harness cannot serve directly as supervision for the target one. We introduce Harness-Zero, which enables harness distillation through agent-as-har…
- Kategoria arxiv
- cs.AI
- Kategoria
- Sztuczna inteligencja