(github.com)

536 points BlueFalconHD | 2 comments | 06 Jul 25 19:50 UTC | HN request time: 0.439s | 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.

1. extraduder_ire ◴[07 Jul 25 08:45 UTC] No.44488115[source]▶

>>44483485 (OP) #

This reminds me of the extensive list of regexes twitch had for filtering allowed usernames that came out when they were hacked.

replies(1): >>44500182 #

2. efilife ◴[08 Jul 25 14:12 UTC] No.44500182[source]▶

>>44488115 (TP) #

I had no idea about this, where can I read them?

↑

I extracted the safety filters from Apple Intelligence models