原书出处
Designing Large Language Model Applications
Pointwise ranking Each candidate document is fed separately to the LLM. The LLM provides a Boolean judgment on its relevance. Alternatively, it can also provide a numerical score, although this is much less reliable. Pairwise ranking For each candidate document pair, the LLM indicates which document is more relevant. To get a complete ranking, N2 such comparisons need to be made. Listwise ranking All the
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/ch12.html
EPUB SHA-256:1b4d2347fec2f949fcddfbaf4fad597a1ae20887a096cfae64c59a619b4781bf