<?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>推理与检索 on caulif · 好奇，然后动手。</title><link>https://www.caulif.com/categories/%E6%8E%A8%E7%90%86%E4%B8%8E%E6%A3%80%E7%B4%A2/</link><description>Recent content in 推理与检索 on caulif · 好奇，然后动手。</description><generator>Hugo -- 0.154.5</generator><language>zh-CN</language><lastBuildDate>Sun, 18 Oct 2026 10:43:25 +0800</lastBuildDate><atom:link href="https://www.caulif.com/categories/%E6%8E%A8%E7%90%86%E4%B8%8E%E6%A3%80%E7%B4%A2/index.xml" rel="self" type="application/rss+xml"/><item><title>LLM Offloading 与 Fate 论文</title><link>https://www.caulif.com/posts/llm-offloading%E4%B8%8Efate%E8%AE%BA%E6%96%87/</link><pubDate>Wed, 02 Sep 2026 00:00:00 +0800</pubDate><guid>https://www.caulif.com/posts/llm-offloading%E4%B8%8Efate%E8%AE%BA%E6%96%87/</guid><description>记录 Fate 论文对 LLM Offloading 的观察，以及 MoE 卸载领域的几个主流方向。</description></item><item><title>RAG简评</title><link>https://www.caulif.com/posts/rag-jian-ping/</link><pubDate>Mon, 10 Aug 2026 00:00:00 +0800</pubDate><guid>https://www.caulif.com/posts/rag-jian-ping/</guid><description>从向量检索的退潮谈 RAG 的技术主导权变化，以及 Agentic Search 和 Context Engineering 的新位置。</description></item><item><title>DSpark</title><link>https://www.caulif.com/posts/dspark/</link><pubDate>Thu, 09 Jul 2026 00:00:00 +0800</pubDate><guid>https://www.caulif.com/posts/dspark/</guid><description>从 speculative decoding 的验证机制讲起，整理 DeepSeek DSpark 在 draft 和 verify 两个阶段的关键改动。</description></item><item><title>RAG</title><link>https://www.caulif.com/posts/rag/</link><pubDate>Fri, 26 Jun 2026 00:00:00 +0800</pubDate><guid>https://www.caulif.com/posts/rag/</guid><description>整理 RAG 的核心思路、离线索引与在线问答流程，以及 Retrieval-Augmented Generation 论文中的检索器和生成器架构。</description></item><item><title>GGUF与量化</title><link>https://www.caulif.com/posts/gguf%E4%B8%8E%E9%87%8F%E5%8C%96/</link><pubDate>Fri, 30 Jan 2026 13:19:41 +0800</pubDate><guid>https://www.caulif.com/posts/gguf%E4%B8%8E%E9%87%8F%E5%8C%96/</guid><description>从 GGUF 文件结构、内存映射、张量卸载和量化格式入手，整理本地大模型文件与推理优化的基础概念。</description></item></channel></rss>