知识卡片
更新知识与学习行为分开选择
频繁变化、需要引用来源的知识常适合外部检索;固定格式和行为映射可考虑微调。两者可组合,不能把检索与训练当互斥方案,也不能保证事实进入参数后就可准确检索和及时更新。
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
- · 微调
原书出处 · 1 处
- · 查看原书依据
new or updated facts or knowledge to the language model, this is better served with techniques like RAG, which we will explore in Chapters 10 and 12. Fine-tuning is best suited for situations where you need the model to learn a particular input-output mapping, be familiarized to a new textual domain, or exhibit more complex capabilities and behavior.
参考来源
- 位置:《Designing Large Language Model Applications》第06章《微调》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch06.html)
- 结论依据:摘录中关于更新知识与学习行为分开选择的定义、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:new or updated facts or knowledge to the language model, this is better served with techniques like RAG, which we will explore in Chapters 10 and 12. Fine-tuning is best suited for situations where you need the model to learn a particular input-output mapping, be familiarized to a new textual domain, or exhibit more complex capabilities and behavior.