(github.com)

536 points BlueFalconHD | 1 comments | 06 Jul 25 19:50 UTC | HN request time: 0.309s | 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

efitz ◴[06 Jul 25 21:04 UTC] No.44484043[source]▶

>>44483485 (OP) #

I’m going to change my name to “Granular Mango Serpent” just to see what those keywords are for in their safety instructions.

replies(2): >>44484103 #>>44486374 #

1. fouronnes3 ◴[06 Jul 25 21:13 UTC] No.44484103[source]▶

>>44484043 #

Granular Mango Serpent is the new David Meyer.

https://arstechnica.com/information-technology/2024/12/certa...

↑

I extracted the safety filters from Apple Intelligence models