←back to thread

I extracted the safety filters from Apple Intelligence models

(github.com)
534 points BlueFalconHD | 3 comments | 06 Jul 25 19:50 UTC | HN request time: 0.646s | 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
Y_Y ◴[07 Jul 25 10:57 UTC] No.44488888[source]▶
Nice to see that we are protected from talking about these weird old dolls:

https://en.wikipedia.org/wiki/Golliwog

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

replies(2): >>44488921 #>>44494232 #
1. oblio ◴[07 Jul 25 11:01 UTC] No.44488921[source]▶
Well, they're not only weird, they're obviously racist doll.
replies(1): >>44489734 #
2. Ey7NFZ3P0nzAe ◴[07 Jul 25 12:36 UTC] No.44489734[source]▶
I want to be able to talk bad about racist things.
replies(1): >>44491608 #
3. chamomeal ◴[07 Jul 25 15:50 UTC] No.44491608[source]▶
Seems like it’s a slur, as well. So not super surprising that it would be blocked tbh