<?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>Mamba on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/mamba/</link><description>Recent content in Mamba 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 11:00:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/mamba/index.xml" rel="self" type="application/rss+xml"/><item><title>Transformer 架构 2026 最新演进：从 Attention 到 MoE 再到 Mamba</title><link>https://guijiagi.com/posts/transformer-architecture-2026-evolution/</link><pubDate>Sun, 28 Jun 2026 11:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/transformer-architecture-2026-evolution/</guid><description>深入解析Transformer架构在2026年的最新演进，涵盖Attention机制优化、MoE混合专家架构、Mamba状态空间模型三大方向</description></item><item><title>超越 Transformer：Mamba/SSM/RWKV 架构深度对比</title><link>https://guijiagi.com/posts/transformer-alternatives-2026/</link><pubDate>Thu, 25 Jun 2026 12:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/transformer-alternatives-2026/</guid><description>深度解析 Mamba、S4/S6、RWKV 等替代 Transformer 的新一代架构，从状态空间模型到线性注意力，全面对比优劣。</description></item></channel></rss>