←back to thread

I extracted the safety filters from Apple Intelligence models

(github.com)
534 points BlueFalconHD | 1 comments | 06 Jul 25 19:50 UTC | HN request time: 0.205s | 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
1f60c ◴[07 Jul 25 12:17 UTC] No.44489569[source]▶
It's pretty easy to understand why Apple doesn't want its models to reproduce racial slurs, but what’s wrong with "Boris Johnson?"

(See, e.g., here: https://github.com/BlueFalconHD/apple_generative_model_safet...)

replies(5): >>44489646 #>>44489672 #>>44489683 #>>44490425 #>>44490666 #
1. qoez ◴[07 Jul 25 12:30 UTC] No.44489683[source]▶
"Justin Trudeau" too. At least it's somewhat unbiased. Still weird imo.