知识卡片
占位替换与同类替换各有代价
特殊占位符便于表示类型但改变自然文本模式;同类假数据更像原文,却可能引入不一致或新的关联。选择方案时应检查实体共指、上下文与任务影响,不能仅看替换字符串是否合法。
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
- · 预训练数据
原书出处 · 1 处
- · 查看原书依据
Replace with a special token For example, a valid phone number can be replaced by the string <phone number>. Replace with a random token of the same entity type For example, replace the name “Clarietta Richards” with “Natasha Bridges,” or any other name. Replace with a shuffled token Entities detected across the dataset can be shuffled. Remove entire document/data source If the amount of PII
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:摘录中关于占位替换与同类替换各有代价的定义、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:Replace with a special token For example, a valid phone number can be replaced by the string <phone number>. Replace with a random token of the same entity type For example, replace the name “Clarietta Richards” with “Natasha Bridges,” or any other name. Replace with a shuffled token Entities detected across the dataset can be shuffled. Remove entire document/data source If the amount of PII