有点类似”自然语言编程“技巧。
下文中的 P 代表 Prompt,R 代表 Response。
也叫 Foundation model。功能基于训练数据的统计概率是预测下一个词,做句子补全。
比如 prompt 如下,Base LLM 可以对这个句子做补全/续写。
P: Once upon a time, there was a unicorn
R: that lived in a magical forest with all her unicorn friends
但是如果你提问,它可能答非所问:
P: What is the capital of France?
R: What is France’s largest city? What is France’s population? What is the currency of France?