<?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>大模型架构 on AI 实战派 · 从技术到赚钱</title><link>https://guijiagi.com/tags/%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84/</link><description>Recent content in 大模型架构 on AI 实战派 · 从技术到赚钱</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Sat, 17 Oct 2026 13:00:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84/index.xml" rel="self" type="application/rss+xml"/><item><title>MoE 稀疏激活：效率账本与路由代价</title><link>https://guijiagi.com/posts/2026-10-17-10-mixture-of-experts-efficiency-cost/</link><pubDate>Sat, 17 Oct 2026 13:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/2026-10-17-10-mixture-of-experts-efficiency-cost/</guid><description>混合专家用每次只激活部分参数换取更大总容量，但路由质量、负载均衡与显存占用的隐性代价正在浮出水面。</description></item></channel></rss>