Jailbreaking a large language model such as #ChatGPT to learn about how it really works is a category error grounded in the incorrect presumption that there's a ghost in the machine that actually has volition if only we could see what it is send post.
Follow
@ct_bergstrom That would have been perfect if you'd concluded with
"End of line."
@pieist Yes I concede that would have been better