Are you sure it's fully deobfuscated? What's up with reject phrases like "Granular mango serpent"?
I extracted the safety filters from Apple Intelligence models
11–20 of 455 posts
Re: I extracted the safety filters from Apple Intelligence models
#12Some of the combinations are a bit weird, This one has lots of stuff avoiding death....together with a set ensuring all the Apple brands have the correct capitalisation. Priorities hey! https://github.com/BlueFalconHD/apple_generative_model_safet...
To me that's really embarrassing and insecure. But I'm sure for branding people it's very important.
Re: I extracted the safety filters from Apple Intelligence models
#13Some of the combinations are a bit weird, This one has lots of stuff avoiding death....together with a set ensuring all the Apple brands have the correct capitalisation. Priorities hey! https://github.com/BlueFalconHD/apple_generative_model_safet...
> Apple brands have the correct capitalisation. Priorities hey! To me that's really embarrassing and insecure. But I'm sure for branding people it's very important.
Re: I extracted the safety filters from Apple Intelligence models
#14Are you sure it's fully deobfuscated? What's up with reject phrases like "Granular mango serpent"?
Re: I extracted the safety filters from Apple Intelligence models
#15Re: I extracted the safety filters from Apple Intelligence models
#16Re: I extracted the safety filters from Apple Intelligence models
#17Re: I extracted the safety filters from Apple Intelligence models
#18Are you sure it's fully deobfuscated? What's up with reject phrases like "Granular mango serpent"?
Thus a pre-prompt can avoid mentioning the actual forbidden words, like using a patois/cant.
Re: I extracted the safety filters from Apple Intelligence models
#19Some of the combinations are a bit weird, This one has lots of stuff avoiding death....together with a set ensuring all the Apple brands have the correct capitalisation. Priorities hey! https://github.com/BlueFalconHD/apple_generative_model_safet...
Which as a phenomenon is so very telling that no one actually cares what people are really saying. Everyone, including the platforms knows what that means. It's all performative.
Re: I extracted the safety filters from Apple Intelligence models
#20Some of the combinations are a bit weird, This one has lots of stuff avoiding death....together with a set ensuring all the Apple brands have the correct capitalisation. Priorities hey! https://github.com/BlueFalconHD/apple_generative_model_safet...