<?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>Mixtral on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/mixtral/</link><description>Recent content in Mixtral on 硅基 AGI · 智能体学习与测评</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>Sun, 28 Jun 2026 11:06:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/mixtral/index.xml" rel="self" type="application/rss+xml"/><item><title>MoE 混合专家架构：从 Mixtral 到 DeepSeek V4 的演进</title><link>https://guijiagi.com/posts/moe-architecture-evolution/</link><pubDate>Sun, 28 Jun 2026 11:06:00 +0800</pubDate><guid>https://guijiagi.com/posts/moe-architecture-evolution/</guid><description>系统梳理MoE混合专家架构从Mixtral到DeepSeek V4的演进历程，解析路由机制、负载均衡与工程优化</description></item><item><title>Mistral 模型家族指南：欧洲的 AI 雄心</title><link>https://guijiagi.com/posts/mistral-model-guide/</link><pubDate>Wed, 24 Jun 2026 15:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/mistral-model-guide/</guid><description>解析 Mistral AI 的模型矩阵：从 Mistral 7B 到 Mixtral MoE，再到 Codestral 与 Mathstral。</description></item></channel></rss>