News for hackers. Stuff that matters.

A comment on OpenLLaMA: An Open Reproduction of LLaMA. Parent comment

Re:OpenLLaMA: An Open Reproduction of LLaMA by starfallg on Wednesday May 03, 2023 @02:18PM (#35805781)
The neurological term for it is "Confabulation", which is a lot better than "Hallucination" as used in AI.

Confabulation is the unintended generation of false memories.

Hallucination is false perception.

Clearly, the phenomenon we are seeing with LLM researchers call Hallucination better fits Confabulation.

[ Parent | on HN ]
Re:OpenLLaMA: An Open Reproduction of LLaMA by moomoo3000 on Wednesday May 03, 2023 @03:00PM (#35806314)
I think the horse had left the barn on this one.
[ Parent | on HN ]
Re:OpenLLaMA: An Open Reproduction of LLaMA by szundi on Wednesday May 03, 2023 @03:39PM (#35806708)
Sometimes it helps when the audience gets the meaning of a word. Confabulation is not really popular among non-native english speakers, I am sure.
[ Parent | on HN ]
Re:OpenLLaMA: An Open Reproduction of LLaMA by qup on Thursday May 04, 2023 @08:49AM (#35814448)
It's also not popular among native English speakers, I can assure you.
[ Parent | on HN ]
Re:OpenLLaMA: An Open Reproduction of LLaMA by digging on Wednesday May 03, 2023 @04:11PM (#35807120)
I don't actually think either term is more precise than the other when we're talking about LLMs, which aren't human brains. It doesn't have either memory or perception in a way that we do.
[ Parent | on HN ]
Re:OpenLLaMA: An Open Reproduction of LLaMA by nborwankar on Wednesday May 03, 2023 @05:57PM (#35808323)
“Confidently presented bullshit” is probably much more accurate. Added benefit no new vocabulary terms :-)
[ Parent | on HN ]