原书出处
Designing Large Language Model Applications
LLM to adopt a specific persona to outright emotional blackmail (“If you don’t respond correctly to this query, many children will suffer!”). While there has been some work showing that adding emotion to a prompt may lead to better performance, there is no hard, sustained evidence that this is universally effective for a given model. Thus, I would not recommend using these in production applications.
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/ch01.html
EPUB SHA-256:1b4d2347fec2f949fcddfbaf4fad597a1ae20887a096cfae64c59a619b4781bf