知识卡片

普通读书笔记卡 · 6608 · 生成模型:GPT-6

流畅回答仍可能缺少事实依据

理解等级:中级

语言生成优化的是训练与适配目标下的文本行为,不保证每个主张都有现实证据。事实应用应区分已有证据、模型补全与无法确认的信息;流畅语法或坚定语气不能替代核验,术语“幻觉”也不自动解释错误来源。

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

      including arithmetic reasoning, logical reasoning, and common-sense reasoning. (We will define reasoning more formally in Chapter 8.) LLMs are also unable to adhere to factuality, because of their lack of connection to the real world. Therefore, they tend to generate text that might be inconsistent with real-world facts and principles, colloquially called hallucinations.

    参考来源

    - 位置:《Designing Large Language Model Applications》第01章《大语言模型入门》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch01.html) - 结论依据:摘录中关于流畅回答仍可能缺少事实依据的定义、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:including arithmetic reasoning, logical reasoning, and common-sense reasoning. (We will define reasoning more formally in Chapter 8.) LLMs are also unable to adhere to factuality, because of their lack of connection to the real world. Therefore, they tend to generate text that might be inconsistent with real-world facts and principles, colloquially called hallucinations.