<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>长时任务 on MessageDaily</title><link>https://inkeast.github.io/MessageDaily/tags/%E9%95%BF%E6%97%B6%E4%BB%BB%E5%8A%A1/</link><description>Recent content in 长时任务 on MessageDaily</description><generator>Hugo</generator><language>zh-cn</language><lastBuildDate>Fri, 12 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://inkeast.github.io/MessageDaily/tags/%E9%95%BF%E6%97%B6%E4%BB%BB%E5%8A%A1/index.xml" rel="self" type="application/rss+xml"/><item><title>SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research 精读</title><link>https://inkeast.github.io/MessageDaily/posts/2026-06-12-searchswarm-paper-reading/</link><pubDate>Fri, 12 Jun 2026 00:00:00 +0000</pubDate><guid>https://inkeast.github.io/MessageDaily/posts/2026-06-12-searchswarm-paper-reading/</guid><description>深度精读 SearchSwarm——首个系统探索如何让 Agent 学会「委派」的工作。论文设计了精巧的 Harness 引导主 Agent 将子任务分派给子 Agent，用合成的轨迹数据通过 SFT 将「委派智能」内化到模型权重中。30B 参数的小模型在 BrowseComp、GAIA 等四个基准上达到同规模最佳，甚至超越 10 倍参数的模型。更令人惊喜的是，委派训练的智能还能泛化到单 Agent 设置和开放式研究任务。</description></item></channel></rss>