SSR：通过逐级规模竞速实现智能体 harness 演化
原标题：Forging More Tasks into Better Harnesses: Agentic Harness Evolution via Successive Scale Racing
AI 摘要
论文提出 Successive Scale Racing（SSR）框架，用于智能体 harness 的自动演化。作者发现，现有方法在扩大演化任务池时性能反而下降，原因是优化过程中证据选择不佳；SSR 通过按质量与覆盖度排序证据组合，并在小算力预算下对候选 harness 进行渐进式竞速筛选来解决该问题。在 SWE-Bench Pro Verified、DeepSWE 和 GAIA-2 三个基准的六种配置上，SSR 均取得最优结果；任务池从 40 增至 200 时通过率从 64.3% 升至 68.3%，同时验证执行次数减少 60%。
正文节选
Forging More Tasks into Better Harnesses: Agentic Harness Evolution via Successive Scale Racing Abstract Agentic harness evolution optimizes an agent’s instructions, tools, and control flow by using its task execution experiences as evolving evidence. While more evolution tasks should intuitively improve performance, our preliminary study reveals a paradoxical failure mode: the performance of existing methods degrades as the task pool expands. Our further analysis demonstrates that this degradat