6 ms·
It's also much slower. LLMs are generating text token at a time. That's not very good for search. Pre-search tokenization however, probably a good fit for LLMs
by coding123 3y ago
It's also much slower. LLMs are generating text token at a time. That's not very good for search.
Pre-search tokenization however, probably a good fit for LLMs.