原书出处
Designing Large Language Model Applications
For the latter, we can control parallel generation through the prompt. For example, say you are writing an article about a tourist site, containing sections like Food, Stay, Safety Tips, etc. You can prompt the LLM to list the sections, marked with special tokens. These sections can then be generated in parallel, assuming the sections are fully independent of each other.
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