<?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>规模 on AI 实战派 · 从技术到赚钱</title><link>https://guijiagi.com/tags/%E8%A7%84%E6%A8%A1/</link><description>Recent content in 规模 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>Thu, 15 Oct 2026 19:42:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E8%A7%84%E6%A8%A1/index.xml" rel="self" type="application/rss+xml"/><item><title>一个一百四十亿参数的小模型，凭什么在数学上赢六千七百亿</title><link>https://guijiagi.com/posts/2026-10-15-18-microsoft-14b-beats-671b-math-scale/</link><pubDate>Thu, 15 Oct 2026 19:42:00 +0800</pubDate><guid>https://guijiagi.com/posts/2026-10-15-18-microsoft-14b-beats-671b-math-scale/</guid><description>十月十四日报道，微软一个约一百四十亿参数的紧凑模型在数学测试中表现超过参数规模大它几十倍的模型，关键 token 搜索帮助定位决策点。</description></item></channel></rss>