知识卡片

普通读书笔记卡 · 6770.a · 生成模型:GPT-6

依据约束回答仍需核实资料

理解等级:中级

Grounding将回答与可查事实、来源或用户上下文联系起来,可用非结构化文档或结构化知识。抽取到一条边只表示系统从材料得到某个主张,材料可能错误或过期;不能把图谱作为可靠事实核查的唯一必要方法。

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

      Grounding is the process of anchoring an LLM’s response in real-world knowledge. This can be via relevant facts and sources or user-relevant information. In the previous chapter, you learned one way to ground LLMs: retrieving and providing them with unstructured text from your knowledge base. In this chapter, you’ll see another helpful way to use structured knowledge—knowledge graphs.

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第11章《信息抽取与知识图谱》(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-11.html) - 结论依据:摘录中关于依据约束回答仍需核实资料的定义、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:Grounding is the process of anchoring an LLM’s response in real-world knowledge. This can be via relevant facts and sources or user-relevant information. In the previous chapter, you learned one way to ground LLMs: retrieving and providing them with unstructured text from your knowledge base. In this chapter, you’ll see another helpful way to use structured knowledge—knowledge graphs.