←back to thread

534 points BlueFalconHD | 1 comments | | HN request time: 0.324s | source

I managed to reverse engineer the encryption (refered to as “Obfuscation” in the framework) responsible for managing the safety filters of Apple Intelligence models. I have extracted them into a repository. I encourage you to take a look around.
Show context
trebligdivad ◴[] No.44483981[source]
Some of the combinations are a bit weird, This one has lots of stuff avoiding death....together with a set ensuring all the Apple brands have the correct capitalisation. Priorities hey!

https://github.com/BlueFalconHD/apple_generative_model_safet...

replies(11): >>44483999 #>>44484073 #>>44484095 #>>44484410 #>>44484636 #>>44486072 #>>44487916 #>>44488185 #>>44488279 #>>44488362 #>>44488856 #
grues-dinner ◴[] No.44484073[source]
Interesting that it didn't seem to include "unalive".

Which as a phenomenon is so very telling that no one actually cares what people are really saying. Everyone, including the platforms knows what that means. It's all performative.

replies(11): >>44484164 #>>44484360 #>>44484635 #>>44484665 #>>44485033 #>>44485034 #>>44486246 #>>44487244 #>>44488055 #>>44488114 #>>44500918 #
martin-t ◴[] No.44484665[source]
No-one cares yet.

There's a very scary potential future in which mega-corporations start actually censoring topics they don't like. For all I know the Chinese government is already doing it, there's no reason the British or US one won't follow suit and mandate such censorship. To protect children / defend against terrorists / fight drugs / stop the spread of misinformation, of course.

replies(2): >>44485059 #>>44490303 #
os2warpman ◴[] No.44490303[source]
HN has censorship that makes those apple rules look like anarchy.

Write a spicy comment and a mod will memory-hole it and someone, usually dang, will reply "tHat'S nOt OuR vIsIon FoR hAcKeR nEwS, pLeAsE bE cIvIl" and we all swallow it like a delicious hot cocoa.

If YC can control their product (and hn IS a product) to annihilate any criticism of their activity or (even former) staff, then Apple is perfectly within their rights to make sure Siri doesn't talk about violence.

No, there's no difference.

replies(1): >>44495862 #
1. martin-t ◴[] No.44495862[source]
Do you mean that HN censors topics/comments which it detects based on advanced filters which search for meaning even when people self-censor and use language to avoid simplistic filters like regex?

HN also has a flagging system and some people really, really hate some kind of speech. Usually they get more offended the more visible it is. A single "bad" word - very offensive to them. A phrase which implies someone is of lesser intelligence or acting in bad faith - sometimes gets a pass, sometimes gets reported. But covert actions like lying, using fallacies to argue or systematic downvoting seem to almost never get punished.