← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8705.b · 生成模型:GPT-6

检索对象不局限外部知识

理解等级:中级

检索可服务事实资料、历史会话、少样本示例及工具说明。每类对象的相关性不同:事实需支持回答,示例需帮助任务映射,工具需符合能力需求;不能用同一主题相似度替代所有用途的评价。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Typical RAG Scenarios Now that we have seen why we need RAG, let’s explore where we can utilize it. The four most popular scenarios are: Retrieving external knowledge This is the predominant use case that has seen a lot of success with productionization. As discussed earlier in the chapter, we can use RAG to plug LLM knowledge gaps or to reduce hallucination risk. Retrieving context

    参考来源

    - 位置:《Designing Large Language Model Applications》第12章《检索增强生成》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch12.html) - 结论依据:摘录中关于检索对象不局限外部知识的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Typical RAG Scenarios Now that we have seen why we need RAG, let’s explore where we can utilize it. The four most popular scenarios are: Retrieving external knowledge This is the predominant use case that has seen a lot of success with productionization. As discussed earlier in the chapter, we can use RAG to plug LLM knowledge gaps or to reduce hallucination risk. Retrieving context