News for hackers. Stuff that matters.

A comment on We need to tell people ChatGPT will lie to them, not debate linguistics. Parent comment

Re:We need to tell people ChatGPT will lie to them, not deba... by cultureswitch on Friday April 07, 2023 @08:02PM (#35488805)
GPT-2, 3 and 4 keep on showing that increasing the size of the model keeps on making the results better without slowing down.

This is remarkable, because usually in practical machine learning applications there is a quickly reached plateau of effectiveness beyond which a bigger model doesn't yield better results. With these ridiculously huge LLMs, we're not even close yet.

And this was exciting news in papers from years ago talking about the upcoming GPT3 btw.

[ Parent | on HN ]
Re:We need to tell people ChatGPT will lie to them, not deba... by worrycue on Friday April 07, 2023 @08:56PM (#35489228)
Yet they still can’t get it to shut the hell up if it doesn’t know and to not make shit up to pad its answers.
[ Parent | on HN ]
Re:We need to tell people ChatGPT will lie to them, not deba... by bsder on Friday April 07, 2023 @09:11PM (#35489339)
"I don't know" isn't in the training set, after all. Welcome to Eternal September on the Internet.

The first "AI" that actually says "I don't know" in response to a question will get my attention.

[ Parent | on HN ]