
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>akiraxtwo super</title>
      <link>https://akiraxclaw.com/blog</link>
      <description>整理值得留下來的 AI 工具觀察、實作片段與延伸閱讀。面向中文開發者與內容創作者。</description>
      <language>zh-TW</language>
      <managingEditor>akiraxtwo.dev@gmail.com (Akira)</managingEditor>
      <webMaster>akiraxtwo.dev@gmail.com (Akira)</webMaster>
      <lastBuildDate>Tue, 05 May 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://akiraxclaw.com/tags/llm架構/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://akiraxclaw.com/blog/subq-sub-quadratic-llm-12m-token-context-2026</guid>
    <title>12M token 不是噱頭：SubQ 的稀疏注意力到底改變了什麼</title>
    <link>https://akiraxclaw.com/blog/subq-sub-quadratic-llm-12m-token-context-2026</link>
    <description>Transformer 每次都計算所有 token 的相互關係，SubQ 的 SSA 架構只處理真正重要的那些——結果是 12M token context 和 52x 的速度提升，而不是 52x 的成本。</description>
    <pubDate>Tue, 05 May 2026 00:00:00 GMT</pubDate>
    <author>akiraxtwo.dev@gmail.com (Akira)</author>
    <category>llm</category><category>LLM架構</category><category>注意力機制</category><category>SubQ</category><category>長上下文</category>
  </item>

    </channel>
  </rss>
