知识卡片
归纳结论保留样本外不确定性
归纳从有限观察推广一般规律,结论强度依赖样本代表性与证据。见过许多圆井盖只能支持常见性判断,不能推出所有井盖必然圆形;额外反例应改变推广范围或置信程度。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 对齐训练与推理
原书出处 · 1 处
- · 查看原书依据
Inductive Reasoning Inductive reasoning involves making generalizations based on a set of observations. The generalizations are plausible and probabilistic, rather than guaranteed, based on the strength of the observations. As an example, upon observing hundreds or even thousands of round manhole covers, one can conclude that manhole covers are generally round. This is not guaranteed to be true, as there might be cities with different
参考来源
- 位置:《Designing Large Language Model Applications》第08章《对齐训练与推理》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch08.html)
- 结论依据:摘录中关于归纳结论保留样本外不确定性的定义、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:Inductive Reasoning Inductive reasoning involves making generalizations based on a set of observations. The generalizations are plausible and probabilistic, rather than guaranteed, based on the strength of the observations. As an example, upon observing hundreds or even thousands of round manhole covers, one can conclude that manhole covers are generally round. This is not guaranteed to be true, as there might be cities with different