(github.com)

534 points BlueFalconHD | 1 comments | 06 Jul 25 19:50 UTC | HN request time: 0.291s | 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. waterproof ◴[07 Jul 25 03:31 UTC] No.44486497[source]▶

>>44483485 (OP) #

Here's a combined file of all the non-locale-specific rules, for easier review: https://github.com/BlueFalconHD/apple_generative_model_safet...

It was generated as part of this PR to consolidate the metadata.json files: https://github.com/BlueFalconHD/apple_generative_model_safet...

↑

I extracted the safety filters from Apple Intelligence models