← 返回卡片

原书出处

Designing Large Language Model Applications

First Edition · 核验版本的摘录快照 · 导入于 2026-11-14T13:54:13.944Z

Unsupervised generation In this technique, the teacher is prompted with instructions and/or examples for solving a task. The teacher’s responses comprise the distillation dataset. This technique is commonly used to teach capabilities like CoT or instruction-following to smaller models. To accomplish that, teacher models are asked to respond to queries with the thought process leading up to the answer. Data augmentation In this technique, the

此预览只包含定位所需的摘录和上下文。

原书定位详情

来源:a19c7933-4eda-4176-aebd-30fe7e1ad292

正文位置:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch09.html

EPUB SHA-256:1b4d2347fec2f949fcddfbaf4fad597a1ae20887a096cfae64c59a619b4781bf