跳到正文
RadarAI· X:Rohan Paul (@rohanpaul_ai)·· 7 小时前AI 评分58

Google 论文提出可自我改进的智能体 harness 框架,门控模块编辑准确率接近全量重写但 token 成本仅约十分之一

Google 论文:智能体 harness 自我改进时门控模块编辑准确率接近全量重写但 token 成本仅约十分之一

AI 导读

Google 论文《Harnesses That Improve Themselves》提出可自我改进的智能体 harness 统一框架,每次只改 4 个可编辑组件中的 1 个,并通过验证集和 token 上限的门控机制决定是否保留。

正文 · 原文

Rohan Paul· @rohanpaul_ai · X·2026-11-20 05:05· 59 分钟前AI 评分54

AI 导读

Google 论文《Harnesses That Improve Themselves》提出可自我改进的智能体 harness 统一框架,核心做法是每次只改 4 个可编辑组件中的 1 个,并通过验证集和 token 上限的门控机制决定是否保留。

正文 · 原文

New Google paper finds that an unchecked full rewrite of an agent harness matched accuracy but cost nearly 10 times more tokens per task than gated modular edits.

that self-improving agent harnesses work best when they change 1 code module at a time and must pass validation.

The harness is the code around a model that manages context, runs tools, and checks work before stopping. A harness tuned for 1 model often stalls or wastes tokens on another.

They let an optimizer model improve a 142-line coding harness by editing 1 of its 4 parts at a time. Each edit had to pass a validation set and a token cap, or it was rejected.

On 40 held-out SWE-bench Lite tasks with Gemini 3.8 Flash, the gated harness solved 40.0% against 5.0% for the seed, using 46.7k tokens per task. A full unchecked rewrite solved 42.5% but used 451.6k tokens, and prompt rules dropped accuracy to 2.5%.

If you let agents tune their harness, gate each change on held-out tasks and token cost, and enforce rules in code, not prompts.

来源:Rohan Paul · x.com

来源:RadarAI · aihot.news