HarnessMM: Recursive Self-Improvement for Harness Token Efficiency· HarnessMM：递归自我改进提升代币效率
Agent harnesses repeatedly invoke reasoning after tool feedback, even when an earlier call has already fixed the next action. We introduce HarnessMM (Harness Minus Minus), a method for improving harness token efficiency through guarded decision reuse and recursive self-improvement (RSI). RSI iteratively refines companion harness configurations from training feedback, with validation selecting the configuration for deployment. A controller at the model-tool boundary reuses resolved decisions to replace later reasoning callbacks while actor weights remain fixed. Executable Decision Graphs store exact actions, literal arguments, causal dependencies, and public guards. Guarded Decision Reuse activates these persistent decisions from observed receipts; Certified Callback Substitution executes t
通过递归自我改进优化Agent调用效率。
- 核心方法
- 引入HarnessMM方法，利用受到保护的决策重用和递归自我改进技术，优化代币使用效率。
- 适合谁读
- 研究者 / 工程师
- 要解决的问题
- 现有Agent框架在工具反馈后重复调用推理，即使早期调用已解决后续动作需求。
- 关键实验
- 通过对比实验展示了HarnessMM在不同任务中对代币效率的具体提升，未提供具体数值。
- 主要贡献
- 提出了Executable Decision Graphs和Guarded Decision Reuse机制，有效减少了无效调用，提高了决策效率。
- 意义与局限
- 本研究可以显著提升多轮推理任务中资源利用效率，但局限于特定代理框架中的应用。