News for hackers. Stuff that matters.
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by seydor on Wednesday December 14, 2022 @05:55PM (#33991958)
Not sure about all its filters being RL. Sometimes it seems to flag its output as inappropriate because of a single word (or none at all). Also it has asymmetric behavior, e.g. it will make a joke about men but refuse to make one about women
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by geraldwhen on Wednesday December 14, 2022 @07:21PM (#33993039)
Competitors will remove the political correctness filters and become more prominent.

A Chinese stable diffusion art site is already gaining traction, no filters.

Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by thedataangel on Wednesday December 14, 2022 @08:54PM (#33994023)
No filters you say? That sounds terrible. Could you provide a link so that I can definitely avoid it.
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by matheusmoreira on Wednesday December 14, 2022 @10:58PM (#33995026)
Hope one day we have the ability to run these things on our own computers with 0 restrictions.
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by astrange on Thursday December 15, 2022 @07:44AM (#33998644)
Already do. (Is the Chinese site actually based on Stable Diffusion though? It looks pretty different.)
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by astrange on Wednesday December 14, 2022 @11:35PM (#33995245)
Until you try to sell your service to women and they leave because it generates porn of them every time they try to make a selfie. Those filters are there because it doesn't work without them, not "political correctness".

(Already happens with Lensa.)

Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by Thorentis on Thursday December 15, 2022 @03:58AM (#33996921)
Link?
Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by Valmar on Wednesday December 14, 2022 @08:58PM (#33994066)
> Not sure about all its filters being RL. Sometimes it seems to flag its output as inappropriate because of a single word (or none at all). Also it has asymmetric behavior, e.g. it will make a joke about men but refuse to make one about women

Probably because in today's age, making jokes about men is a-okay, but making jokes against women might be perceived as misogynist. Potential misandry is okay, by comparison.

Re:How Does GPT Obtain Its Ability? Tracing Emergent Abiliti... by fullsend on Wednesday December 14, 2022 @11:30PM (#33995210)
That’s wrong. If you can make it do so, it will make jokes about anyone. See the example “write a function to determine if an engineer is good at their job” and you can get it to return a terrible Python joke of “if gender == male return True else…”.