News for hackers. Stuff that matters.

A comment on OpenAI's plans according to sama. Parent comment

Re:OpenAI's plans according to sama by sebzim4500 on Wednesday May 31, 2023 @03:17PM (#36142742)
It is weird, but GPT-3 is worse than much smaller LLaMA models so I doubt it would see much use anyway.
[ Parent | on HN ]
Re:OpenAI's plans according to sama by killjoywashere on Wednesday May 31, 2023 @03:21PM (#36142810)
How do you measure this? Pointers to papers would be very helpful
[ Parent | on HN ]
Re:OpenAI's plans according to sama by sebzim4500 on Wednesday May 31, 2023 @03:22PM (#36142823)
The LLaMA paper had a bunch of comparisons
[ Parent | on HN ]
Re:OpenAI's plans according to sama by colordrops on Thursday June 01, 2023 @03:17AM (#36148375)
Aren't the LLaMA weights leaked though? Did Facebook ever open up its license?
[ Parent | on HN ]
Re:OpenAI's plans according to sama by renonce on Friday June 02, 2023 @08:07AM (#36164032)
Doesn’t matter if you only use it yourself. No one will know.
[ Parent | on HN ]
Re:OpenAI's plans according to sama by flangola7 on Wednesday May 31, 2023 @03:20PM (#36142790)
Are you referring to DaVinci or ChatGPT-3.5
[ Parent | on HN ]
Re:OpenAI's plans according to sama by sebzim4500 on Wednesday May 31, 2023 @03:21PM (#36142803)
DaVinci
[ Parent | on HN ]