Northwestern 与 Google 论文提出 Harness2：无需答案键自动改进智能体 harness

Rohan Paul· @rohanpaul_ai · X·2026-11-15 08:08· 1 小时前AI 评分53

AI 导读

Northwestern 与 Google 的论文 Harness² 提出 plug-and-play 的递归 harnessing 框架，让智能体 harness（提示词、工具、技能与记忆的封装）在新任务上无答案键自我改进：编辑模型提出修改，重跑任务只保留有效的修改，经验存入库指导后续任务。

正文 · 原文

New Northwestern and Google paper shows that agent harnesses can improve themselves on new tasks without answer keys, by test-running edits and keeping what helps.

An agent's harness is the prompts, tools, skills, and memory wrapped around the model. Tuning it by hand is slow, and new real-world tasks come with no answer key to tune against.

Harness2 runs the agent, has an editor model propose harness fixes, then reruns the task to see which fixes improved the agent's behavior. Lessons from each task go into a library that guides edits on the next one.

A single improvement step raised GLM-5's WorkBuddy-Bench score by 11.5 points, past Gemini 3.5 Flash on its stock harness. Editing the harness without those test runs instead dropped coding scores from 73.4 to 63.2.

Treat your agent's harness as something to revise per task, and keep only the edits that change its behavior on real runs.

来源：Rohan Paul · x.com