<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Looped Transformer on AI内参</title>
    <link>https://www.neican.ai/tags/looped-transformer/</link>
    <description>Recent content in Looped Transformer on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Thu, 10 Sep 2026 19:52:00 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/looped-transformer/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>GPT-6 Astra 发布：ARC-AGI-3 拿到 99.9% 之后，looped transformer 成了新的架构焦点</title>
      <link>https://www.neican.ai/insights/gpt6-astra-looped-20260910195200001-0/</link>
      <pubDate>Thu, 10 Sep 2026 19:52:00 +0800</pubDate>
      <guid>https://www.neican.ai/insights/gpt6-astra-looped-20260910195200001-0/</guid>
      <description>OpenAI 正式发布面向专业工作场景的 GPT-6 Astra，已在 ChatGPT Work、Codex 和 API 上线。Sebastian Raschka 的评测显示它在 ARC-AGI-3 上拿到 99.9%（前代仅 7.8%），并从架构视角拆解了 looped transformer 的原理——同时给出独立判断：循环架构与“隐藏思维链”传闻无关。</description>
    </item>
  </channel>
</rss>
