<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[长上下文模型出来以后，RAG 还有必要吗？]]></title><description><![CDATA[<p dir="auto">现在模型上下文这么长，直接把文档塞进去是不是比 RAG 简单？</p>
]]></description><link>https://localaihub.com/topic/67/长上下文模型出来以后-rag-还有必要吗</link><generator>RSS for Node</generator><lastBuildDate>Wed, 03 Jun 2026 18:50:44 GMT</lastBuildDate><atom:link href="https://localaihub.com/topic/67.rss" rel="self" type="application/rss+xml"/><pubDate>Sun, 03 May 2026 16:06:00 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 07:09:00 GMT]]></title><description><![CDATA[<p dir="auto">这句比较准。生产系统通常是混合，不是单一招式。</p>
]]></description><link>https://localaihub.com/post/288</link><guid isPermaLink="true">https://localaihub.com/post/288</guid><dc:creator><![CDATA[nora]]></dc:creator><pubDate>Mon, 04 May 2026 07:09:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 06:38:00 GMT]]></title><description><![CDATA[<p dir="auto">别把技术路线变宗教。短资料长上下文，长期知识 RAG，关键事实结构化。</p>
]]></description><link>https://localaihub.com/post/287</link><guid isPermaLink="true">https://localaihub.com/post/287</guid><dc:creator><![CDATA[林小北]]></dc:creator><pubDate>Mon, 04 May 2026 06:38:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 05:07:00 GMT]]></title><description><![CDATA[<p dir="auto">对，还有成本和可解释。企业里“为什么这么答”比“能不能答”更重要。</p>
]]></description><link>https://localaihub.com/post/286</link><guid isPermaLink="true">https://localaihub.com/post/286</guid><dc:creator><![CDATA[小风扇]]></dc:creator><pubDate>Mon, 04 May 2026 05:07:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 04:25:00 GMT]]></title><description><![CDATA[<p dir="auto">所以边界是资料规模、更新频率、权限和引用要求？</p>
]]></description><link>https://localaihub.com/post/285</link><guid isPermaLink="true">https://localaihub.com/post/285</guid><dc:creator><![CDATA[小键盘]]></dc:creator><pubDate>Mon, 04 May 2026 04:25:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 03:11:00 GMT]]></title><description><![CDATA[<p dir="auto">我们做会议纪要问答，单会 2 小时转写直接塞效果不错。跨项目知识库还是 RAG。</p>
]]></description><link>https://localaihub.com/post/284</link><guid isPermaLink="true">https://localaihub.com/post/284</guid><dc:creator><![CDATA[米饭]]></dc:creator><pubDate>Mon, 04 May 2026 03:11:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Mon, 04 May 2026 01:00:00 GMT]]></title><description><![CDATA[<p dir="auto">最怕把长上下文当垃圾桶。日志、历史聊天、文档全塞，最后不知道模型依据什么答。</p>
]]></description><link>https://localaihub.com/post/283</link><guid isPermaLink="true">https://localaihub.com/post/283</guid><dc:creator><![CDATA[rootless]]></dc:creator><pubDate>Mon, 04 May 2026 01:00:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 23:18:00 GMT]]></title><description><![CDATA[<p dir="auto">这个我同意。RAG 和长上下文不是二选一，可以检索少量大块。</p>
]]></description><link>https://localaihub.com/post/282</link><guid isPermaLink="true">https://localaihub.com/post/282</guid><dc:creator><![CDATA[小路灯]]></dc:creator><pubDate>Sun, 03 May 2026 23:18:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 23:06:00 GMT]]></title><description><![CDATA[<p dir="auto">我反而觉得长上下文能减少切块复杂度。先检索章节，再给大段上下文。</p>
]]></description><link>https://localaihub.com/post/281</link><guid isPermaLink="true">https://localaihub.com/post/281</guid><dc:creator><![CDATA[半糖]]></dc:creator><pubDate>Sun, 03 May 2026 23:06:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 22:38:00 GMT]]></title><description><![CDATA[<p dir="auto">还有权限。你不能为了省 RAG，把用户无权看的文档也一起塞给模型。</p>
]]></description><link>https://localaihub.com/post/280</link><guid isPermaLink="true">https://localaihub.com/post/280</guid><dc:creator><![CDATA[MingK]]></dc:creator><pubDate>Sun, 03 May 2026 22:38:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 19:46:00 GMT]]></title><description><![CDATA[<p dir="auto">长上下文成本也要算。每次把大文档塞进去，延迟和费用都上去。</p>
]]></description><link>https://localaihub.com/post/279</link><guid isPermaLink="true">https://localaihub.com/post/279</guid><dc:creator><![CDATA[阿航]]></dc:creator><pubDate>Sun, 03 May 2026 19:46:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 18:59:00 GMT]]></title><description><![CDATA[<p dir="auto">Lost in the Middle 那篇提醒过，长上下文里信息位置也会影响模型使用。不是塞进去就等于读懂。</p>
]]></description><link>https://localaihub.com/post/278</link><guid isPermaLink="true">https://localaihub.com/post/278</guid><dc:creator><![CDATA[nora]]></dc:creator><pubDate>Sun, 03 May 2026 18:59:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 17:57:00 GMT]]></title><description><![CDATA[<p dir="auto">是坑，所以要看场景。10 页项目说明直接塞上下文可以；几千份制度不行。</p>
]]></description><link>https://localaihub.com/post/277</link><guid isPermaLink="true">https://localaihub.com/post/277</guid><dc:creator><![CDATA[林小北]]></dc:creator><pubDate>Sun, 03 May 2026 17:57:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 16:44:00 GMT]]></title><description><![CDATA[<p dir="auto">但 RAG 调参很烦，切块、向量库、rerank 都是坑。</p>
]]></description><link>https://localaihub.com/post/276</link><guid isPermaLink="true">https://localaihub.com/post/276</guid><dc:creator><![CDATA[小树]]></dc:creator><pubDate>Sun, 03 May 2026 16:44:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型出来以后，RAG 还有必要吗？ on Sun, 03 May 2026 16:17:00 GMT]]></title><description><![CDATA[<p dir="auto">简单不等于可靠。长上下文解决“放得下”，没解决“找得准”和“证据可控”。</p>
]]></description><link>https://localaihub.com/post/275</link><guid isPermaLink="true">https://localaihub.com/post/275</guid><dc:creator><![CDATA[小风扇]]></dc:creator><pubDate>Sun, 03 May 2026 16:17:00 GMT</pubDate></item></channel></rss>