<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>状态追踪 on AI内参</title>
    <link>https://www.neican.ai/tags/%E7%8A%B6%E6%80%81%E8%BF%BD%E8%B8%AA/</link>
    <description>Recent content in 状态追踪 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Wed, 17 Jun 2026 21:10:06 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/%E7%8A%B6%E6%80%81%E8%BF%BD%E8%B8%AA/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>超越思维链的囚徒困境：Transformer拓扑缺陷与循环架构的回归</title>
      <link>https://www.neican.ai/insights/transformer-20260617211006323-0/</link>
      <pubDate>Wed, 17 Jun 2026 21:10:06 +0800</pubDate>
      <guid>https://www.neican.ai/insights/transformer-20260617211006323-0/</guid>
      <description>DeepMind研究揭示Transformer架构因拓扑限制导致状态追踪缺失，思维链是昂贵的补丁。未来架构将回归循环机制，实现从“历史检索”向“动态认知”的范式转变，这对降低推理成本与提升智能体连贯性至关重要。</description>
    </item>
    <item>
      <title>大语言模型如何被一场古老棋局“考倒”：ChatGPT与“理解”的边界</title>
      <link>https://www.neican.ai/insights/chatgpt-20250617202000390-6/</link>
      <pubDate>Tue, 17 Jun 2025 20:20:00 +0800</pubDate>
      <guid>https://www.neican.ai/insights/chatgpt-20250617202000390-6/</guid>
      <description>一场ChatGPT与1979年《Video Chess》的对局以大语言模型惨败告终，暴露了其在处理离散、规则严格的状态追踪任务上的固有弱点。此次事件引发了对当前AI能力，尤其是LLM“理解”边界的深刻反思，提醒业界和公众需更清醒地认识到AI的局限性，并呼吁构建更符合任务需求的混合AI系统。</description>
    </item>
  </channel>
</rss>
