logoalt Hacker News

krapp • yesterday at 11:58 PM • 1 reply • view on HN

Yes. And none of them are reliable. LLM "hallucinations" are implicit to their structure and function, you can't just tell them to be honest.

inb4 "humans aren't reliable either." You, person typing that very response out right now, you are part of the problem.


Replies

groundzeros2015 • today at 12:46 AM

that’s a general complain about LLMs not this application

➕ show 1 reply