1
Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts
新方法让AI智能体通过自我偏好优化轨迹，无需人工标注数据即可持续提升能力，为Agent自我进化开辟新思路。
arXiv:2606.05922v1 Announce Type: new Abstract: AI agents rely on a harness of skills, tools, and workflows to solve complex problems. Continually imp…