Hacker News new | ask | show | jobs
by wkat4242 639 days ago
Except we will never solve that for LLMs. It's just how they work. LLM output will always be a "best guess"