<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>NousResearch on AI内参</title>
    <link>https://www.neican.ai/tags/nousresearch/</link>
    <description>Recent content in NousResearch on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Fri, 15 May 2026 21:10:03 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/nousresearch/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>超越算力堆叠：从“Token叠加”看大模型预训练的“粗读”范式革命</title>
      <link>https://www.neican.ai/insights/token-20260515211003597-0/</link>
      <pubDate>Fri, 15 May 2026 21:10:03 +0800</pubDate>
      <guid>https://www.neican.ai/insights/token-20260515211003597-0/</guid>
      <description>TST 方法通过在预训练初期引入“词元叠加”策略，成功将算力成本大幅压缩，标志着大模型训练正在从“堆砌算力”转向“优化学习路径”的精细化时代。这一创新不仅降低了技术研发门槛，也为未来 AI 规模化应用提供了更具性价比的工程路径。</description>
    </item>
  </channel>
</rss>
