News for hackers. Stuff that matters.

A comment on OpenAI employee: GPT-4.5 rumor was a hallucination. Parent comment

Re:OpenAI employee: GPT-4.5 rumor was a hallucination by arthurcolle on Sunday December 17, 2023 @05:55PM (#38677259)
Great theory, this makes a lot of sense. It looks like they already RLHF'ed (I keep using this as a verb, but maybe its really just instruction tuning?) this behavior out of ChatGPT! https://news.ycombinator.com/item?id=38677025 Really fast, I wonder how this works. Anyone have any ideas / knowledge on how they deploy little incremental fixes to exploited jailbreaks, etc?)
[ Parent | on HN ]
Re:OpenAI employee: GPT-4.5 rumor was a hallucination by krisoft on Sunday December 17, 2023 @06:02PM (#38677299)
> Anyone have any ideas / knowledge on how they deploy little incremental fixes to exploited jailbreaks, etc?

LoRa[1] would be my guess.

For detailed explanation I recommend the paper. But the short explanation is that it is a trick which lets you train a smaller, lower dimensional model which when you add to the original model it gets you the result you want.

1: https://arxiv.org/abs/2106.09685

[ Parent | on HN ]
Re:OpenAI employee: GPT-4.5 rumor was a hallucination by arthurcolle on Sunday December 17, 2023 @07:27PM (#38677887)
is this another expert? can they hot swap in experts and remove others?
[ Parent | on HN ]
Re:OpenAI employee: GPT-4.5 rumor was a hallucination by azeirah on Sunday December 17, 2023 @05:58PM (#38677277)
Not sure about this scenario precisely but I believe when it comes to filtering certain responses they have a separate model running in a similar manner to what meta recently released. I believe it's called llama purple?

https://ai.meta.com/llama/purple-llama/

[ Parent | on HN ]