2 ms·
I just tested this using Calude and at least with 4.5 this does not seem to be possible. The context grows very quickly and the LLM gets lost and starts halluci
by throwa356262 8mo ago
I just tested this using Calude and at least with 4.5 this does not seem to be possible. The context grows very quickly and the LLM gets lost and starts hallucinating. Maybe I am missing some key ingredient here?
Of course, if you have large team of AI and security experts and an unlimited token budget things can look different.