<?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[长上下文模型是不是可以不用 RAG 了]]></title><description><![CDATA[<p dir="auto">现在好多模型都支持很长上下文了，企业知识库是不是可以直接把文档塞进去，不做 RAG？</p>
]]></description><link>https://localaihub.com/topic/186/长上下文模型是不是可以不用-rag-了</link><generator>RSS for Node</generator><lastBuildDate>Wed, 03 Jun 2026 19:20:35 GMT</lastBuildDate><atom:link href="https://localaihub.com/topic/186.rss" rel="self" type="application/rss+xml"/><pubDate>Wed, 13 May 2026 12:42:00 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 12:21:00 GMT]]></title><description><![CDATA[<p dir="auto">还有一点：长上下文贵。能用 8k 解决的问题，不要硬上 200k。</p>
]]></description><link>https://localaihub.com/post/2021</link><guid isPermaLink="true">https://localaihub.com/post/2021</guid><dc:creator><![CDATA[阿航]]></dc:creator><pubDate>Thu, 14 May 2026 12:21:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 10:30:00 GMT]]></title><description><![CDATA[<p dir="auto">适合长文总结、合同审阅、代码仓库片段阅读、会议资料综合。前提是资料集合本身可控。</p>
]]></description><link>https://localaihub.com/post/2020</link><guid isPermaLink="true">https://localaihub.com/post/2020</guid><dc:creator><![CDATA[陈一]]></dc:creator><pubDate>Thu, 14 May 2026 10:30:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 09:18:00 GMT]]></title><description><![CDATA[<p dir="auto">那长上下文主要适合什么？</p>
]]></description><link>https://localaihub.com/post/2019</link><guid isPermaLink="true">https://localaihub.com/post/2019</guid><dc:creator><![CDATA[小鼠标]]></dc:creator><pubDate>Thu, 14 May 2026 09:18:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 07:12:00 GMT]]></title><description><![CDATA[<p dir="auto">我现在做法是：检索召回一批，rerank 留少量，再把文档标题、更新时间、权限一起进上下文。</p>
]]></description><link>https://localaihub.com/post/2018</link><guid isPermaLink="true">https://localaihub.com/post/2018</guid><dc:creator><![CDATA[melo]]></dc:creator><pubDate>Thu, 14 May 2026 07:12:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 04:17:00 GMT]]></title><description><![CDATA[<p dir="auto">先按任务放。用户问报销流程，就不要把入职、采购、绩效都塞进去。</p>
]]></description><link>https://localaihub.com/post/2017</link><guid isPermaLink="true">https://localaihub.com/post/2017</guid><dc:creator><![CDATA[林小北]]></dc:creator><pubDate>Thu, 14 May 2026 04:17:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Thu, 14 May 2026 01:33:00 GMT]]></title><description><![CDATA[<p dir="auto">那怎么判断放多少？</p>
]]></description><link>https://localaihub.com/post/2016</link><guid isPermaLink="true">https://localaihub.com/post/2016</guid><dc:creator><![CDATA[小吴]]></dc:creator><pubDate>Thu, 14 May 2026 01:33:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 23:52:00 GMT]]></title><description><![CDATA[<p dir="auto">这叫上下文污染。不是资料越多越安全，过期资料混进去更危险。</p>
]]></description><link>https://localaihub.com/post/2015</link><guid isPermaLink="true">https://localaihub.com/post/2015</guid><dc:creator><![CDATA[nora]]></dc:creator><pubDate>Wed, 13 May 2026 23:52:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 23:30:00 GMT]]></title><description><![CDATA[<p dir="auto">我们试过把整本手册塞进去，模型会引用旧条款，因为旧条款也在上下文里。</p>
]]></description><link>https://localaihub.com/post/2014</link><guid isPermaLink="true">https://localaihub.com/post/2014</guid><dc:creator><![CDATA[半截薯条]]></dc:creator><pubDate>Wed, 13 May 2026 23:30:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 22:23:00 GMT]]></title><description><![CDATA[<p dir="auto">不是。长上下文只代表窗口更大，不代表模型会稳定关注每个细节。位置、重复、冲突内容都会影响答案。</p>
]]></description><link>https://localaihub.com/post/2013</link><guid isPermaLink="true">https://localaihub.com/post/2013</guid><dc:creator><![CDATA[Grace]]></dc:creator><pubDate>Wed, 13 May 2026 22:23:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 20:25:00 GMT]]></title><description><![CDATA[<p dir="auto">长上下文不是越长越聪明吗？</p>
]]></description><link>https://localaihub.com/post/2012</link><guid isPermaLink="true">https://localaihub.com/post/2012</guid><dc:creator><![CDATA[普通网友A]]></dc:creator><pubDate>Wed, 13 May 2026 20:25:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 19:54:00 GMT]]></title><description><![CDATA[<p dir="auto">所以不是 RAG 必须上，是任务决定。几十页临时合同可以长上下文，几万份制度肯定要检索。</p>
]]></description><link>https://localaihub.com/post/2011</link><guid isPermaLink="true">https://localaihub.com/post/2011</guid><dc:creator><![CDATA[陈一]]></dc:creator><pubDate>Wed, 13 May 2026 19:54:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 17:15:00 GMT]]></title><description><![CDATA[<p dir="auto">但 RAG 也很麻烦。切块、向量、rerank、权限，全是坑。</p>
]]></description><link>https://localaihub.com/post/2010</link><guid isPermaLink="true">https://localaihub.com/post/2010</guid><dc:creator><![CDATA[小蓝]]></dc:creator><pubDate>Wed, 13 May 2026 17:15:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 15:41:00 GMT]]></title><description><![CDATA[<p dir="auto">我反对完全不用 RAG。资料多的时候，先检索再喂模型还是必要的，不然成本和延迟都爆。</p>
]]></description><link>https://localaihub.com/post/2009</link><guid isPermaLink="true">https://localaihub.com/post/2009</guid><dc:creator><![CDATA[小风扇]]></dc:creator><pubDate>Wed, 13 May 2026 15:41:00 GMT</pubDate></item><item><title><![CDATA[Reply to 长上下文模型是不是可以不用 RAG 了 on Wed, 13 May 2026 14:59:00 GMT]]></title><description><![CDATA[<p dir="auto">先别把长上下文当垃圾桶。它解决“能放多少”，没解决“该放什么”和“怎么验证来源”。</p>
]]></description><link>https://localaihub.com/post/2008</link><guid isPermaLink="true">https://localaihub.com/post/2008</guid><dc:creator><![CDATA[阿航]]></dc:creator><pubDate>Wed, 13 May 2026 14:59:00 GMT</pubDate></item></channel></rss>