<?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>基础 on Latent</title>
    <link>https://latent-site.pages.dev/tags/%E5%9F%BA%E7%A1%80/</link>
    <description>Recent content in 基础 on Latent</description>
    <generator>Hugo</generator>
    <language>zh</language>
    <copyright>2026 Latent</copyright>
    <lastBuildDate>Sun, 05 Jul 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://latent-site.pages.dev/tags/%E5%9F%BA%E7%A1%80/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>01 | 大模型是怎么工作的——从 Transformer 到 Tokenization</title>
      <link>https://latent-site.pages.dev/posts/2026/07/01-%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%98%AF%E6%80%8E%E4%B9%88%E5%B7%A5%E4%BD%9C%E7%9A%84%E4%BB%8E-transformer-%E5%88%B0-tokenization/</link>
      <pubDate>Sun, 05 Jul 2026 00:00:00 +0000</pubDate>
      <guid>https://latent-site.pages.dev/posts/2026/07/01-%E5%A4%A7%E6%A8%A1%E5%9E%8B%E6%98%AF%E6%80%8E%E4%B9%88%E5%B7%A5%E4%BD%9C%E7%9A%84%E4%BB%8E-transformer-%E5%88%B0-tokenization/</guid>
      <description>不讲公式，用直觉和具体例子解释大模型的核心机制：Transformer 架构、Tokenization 原理、Context Window 概念，以及自回归生成的工作方式。</description>
    </item>
    <item>
      <title>02 | Prompt Engineering——从入门到精通</title>
      <link>https://latent-site.pages.dev/posts/2026/07/02-prompt-engineering%E4%BB%8E%E5%85%A5%E9%97%A8%E5%88%B0%E7%B2%BE%E9%80%9A/</link>
      <pubDate>Sun, 05 Jul 2026 00:00:00 +0000</pubDate>
      <guid>https://latent-site.pages.dev/posts/2026/07/02-prompt-engineering%E4%BB%8E%E5%85%A5%E9%97%A8%E5%88%B0%E7%B2%BE%E9%80%9A/</guid>
      <description>系统讲解 Prompt Engineering 的核心技巧：Zero-shot、Few-shot、Chain-of-Thought、Self-Consistency，以及 System Prompt 设计原则。每个技巧配具体示例和使用场景。</description>
    </item>
    <item>
      <title>03 | 推理模型——o1、R1 和 test-time compute 到底在做什么</title>
      <link>https://latent-site.pages.dev/posts/2026/07/03-%E6%8E%A8%E7%90%86%E6%A8%A1%E5%9E%8Bo1r1-%E5%92%8C-test-time-compute-%E5%88%B0%E5%BA%95%E5%9C%A8%E5%81%9A%E4%BB%80%E4%B9%88/</link>
      <pubDate>Sun, 05 Jul 2026 00:00:00 +0000</pubDate>
      <guid>https://latent-site.pages.dev/posts/2026/07/03-%E6%8E%A8%E7%90%86%E6%A8%A1%E5%9E%8Bo1r1-%E5%92%8C-test-time-compute-%E5%88%B0%E5%BA%95%E5%9C%A8%E5%81%9A%E4%BB%80%E4%B9%88/</guid>
      <description>解释推理模型（o1/o3、DeepSeek-R1）的工作原理：test-time compute、内部思维链、强化学习训练，以及它和传统 CoT Prompt 的本质区别。</description>
    </item>
  </channel>
</rss>
