https://jizuiba.netlify.app/categories/rag/Recent content in RAG on Jizuiba`s BlogHugozh-cnSat, 28 Mar 2026 00:00:00 +0000https://jizuiba.netlify.app/posts/rag-%E6%A3%80%E7%B4%A2%E4%BC%98%E5%8C%96/Sat, 28 Mar 2026 00:00:00 +0000https://jizuiba.netlify.app/posts/rag-%E6%A3%80%E7%B4%A2%E4%BC%98%E5%8C%96/<h2 id="检索优化是在干啥">检索优化是在干啥？</h2> <p>基础的 RAG 调用：</p> <figure class="article-codeblock article-codeblock--text" data-code-language="text" style="--code-max-height: 32rem;"><figcaption class="article-codeblock__header" aria-label="Code block tools"><button type="button" class="article-codeblock__copy" data-copy-default="复制" data-copy-success="已复制" data-copy-error="失败" aria-label="复制"> <svg class="article-codeblock__copy-icon" width="14" height="14" viewBox="0 0 24 24" aria-hidden="true"> <rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect> <path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path> </svg> <span class="article-codeblock__copy-text">复制</span> </button><span class="article-codeblock__language" aria-label="Code language">text</span></figcaption><div class="article-codeblock__body"><div class="highlight"><pre tabindex="0" class="chroma"><code class="language-text" data-lang="text"><span class="line"><span class="cl">用户问题 → embedding → 向量检索 → top-k → 给 LLM</span></span></code></pre></div></div> </figure><p>可能存在的问题：</p>https://jizuiba.netlify.app/posts/rag-%E7%9B%B8%E5%85%B3%E7%9A%84%E7%9F%A5%E8%AF%86%E5%90%8D%E8%AF%8D/Fri, 27 Mar 2026 00:00:00 +0000https://jizuiba.netlify.app/posts/rag-%E7%9B%B8%E5%85%B3%E7%9A%84%E7%9F%A5%E8%AF%86%E5%90%8D%E8%AF%8D/<h2 id="chunk分块">Chunk（分块）</h2> <p><strong>是什么</strong>：</p> <p>把长文档切一小段一小段，再做向量化存储。</p> <p><strong>为什么需要它</strong></p> <ul> <li>LLM 的上下文限制</li> <li>检索时需要“精准命中片段”而不是一整片文章</li> </ul> <p><strong>例如</strong>：</p> <p>一篇10000字的文章 -&gt; 切成：</p>