---
title: "Task-CoEvolve：自适应验证任务选择实现高效LLM智能体框架优化"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.20169v1"
permalink: "https://feed.astralor.com/items/itm_5bcda1f61096914851c3e859"
published: "2026-08-20T15:24:54.000Z"
score: 74
channel: "tech"
tags: ["Harness Optimization","LLM Agent","Task-CoEvolve","Terminal-Bench 2.1","东京大学","方差加权采样"]
---

# Task-CoEvolve：自适应验证任务选择实现高效LLM智能体框架优化

## AI 点评

> 这篇论文值得所有做Agent系统优化的工程师关注，它用很轻量的方法解决了重评估的痛点，而且开箱即用。如果你们团队正在反复调harnass代码，这个思路能帮你省下好几倍算力。

## 速览

东京大学团队提出Task-CoEvolve，用于高效优化LLM智能体框架（harness）。传统方法每轮迭代需在完整固定验证集上评估，成本高昂。Task-CoEvolve通过方差加权采样，聚焦于智能体能力边界附近的任务，并基于采样概率估计全集分数，动态适应框架演化。在在线文本分类和Terminal-Bench 2.1上，该方法比固定子集基线更优，与全量搜索最终性能持平，同时将优化期间评估次数减少80%。代码已开源。

## 关键点

- 提出Task-CoEvolve方法，将验证任务与智能体框架协同演化，动态选择信息量大的任务。
- 基于候选框架分歧度进行方差加权采样，聚焦能力边界附近任务，并用采样概率校正实现全集分数估计。
- 在多个基准上达到全量搜索的最终性能，评估次数减少80%，大幅降低优化成本。

## 全文

We present a novel approach to efficient LLM agent harness optimization through adaptive validation task selection. Harness optimization iteratively rewrites the harness code based on validation performance, enabling substantial performance gains without updating the underlying model weights. Existing approaches, however, evaluate a fixed validation set in full at every iteration, incurring substantial evaluation costs even on tasks that become less discriminative as the harness evolves. We propose $\textbf{Task-CoEvolve}$, which co-evolves the validation tasks with the harness by addressing two challenges: selecting informative tasks and estimating full-set performance from partial evaluations. Task-CoEvolve builds on the observation that tasks on which candidate harnesses disagree are more informative for distinguishing among them than tasks that are consistently solved or failed. It uses variance-weighted sampling based on past outcomes to focus evaluation on tasks near the agent's capability frontier, with the sampling distribution adapting as the harness evolves. It then estimates full-set scores from the sampled tasks by accounting for their sampling probabilities, enabling consistent comparisons across iterations despite evaluating different subsets. Experiments on online text classification and Terminal-Bench 2.1 show that Task-CoEvolve consistently outperforms fixed-subset baselines and matches the final performance of full-set search while reducing the number of evaluations during optimization by 80%. Code will be released at https://github.com/Agent4Science-UTokyo/Task-CoEvolve.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_5bcda1f61096914851c3e859