知识卡片

普通读书笔记卡 · 6788.d · 生成模型:GPT-6

知识库没有收录不等于事实错误

理解等级:中级

缺失条目可能是覆盖不足,也可能是问题超出业务范围。可按产品契约回退或说明缺依据,但不应直接宣称该事实为假;RAG还需检查检索文本是否支持回答,而不是只看出现了相似片段。

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

      In chapters 10 and 11, you learned about a technique for controlling what generative models say and helping LLMs be more helpful. You can have the chatbot base its answers on facts in your knowledge base, rather than have it make up its own facts and references. A knowledge graph is especially useful when you want to fact-check the content the LLM generates. If a

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第12章《对话引擎与会话设计》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-12.html) - 结论依据:摘录中关于知识库没有收录不等于事实错误的定义、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:In chapters 10 and 11, you learned about a technique for controlling what generative models say and helping LLMs be more helpful. You can have the chatbot base its answers on facts in your knowledge base, rather than have it make up its own facts and references. A knowledge graph is especially useful when you want to fact-check the content the LLM generates. If a