知识卡片

普通读书笔记卡 · 6616.a · 生成模型:GPT-6

公开数据仍受原始语境约束

理解等级:中级

某信息在个人网页可见,并不代表将其纳入模型后更广泛传播与原共享目的等价。数据处理应考虑可发现性、预期用途和与其他信息的可链接性;隐私判断不能只问它是否曾公开。

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

      shared my PII on my blog, which resides in an obscure corner of the internet and is not easily discoverable through search engines, but if it ends up being added to a pre-training dataset, it suddenly brings this data into the spotlight, without my consent. This concept is called contextual integrity: data should only be shared in the original context in which it was shared.

    参考来源

    - 位置:《Designing Large Language Model Applications》第02章《预训练数据》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch02.html) - 结论依据:摘录中关于公开数据仍受原始语境约束的定义、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:shared my PII on my blog, which resides in an obscure corner of the internet and is not easily discoverable through search engines, but if it ends up being added to a pre-training dataset, it suddenly brings this data into the spotlight, without my consent. This concept is called contextual integrity: data should only be shared in the original context in which it was shared.