Spaces:
Running
on
CPU Upgrade
Ethics filtering - overstepping, overzealous
For now, we'll have to continue with ChatGPT - the Hugging ethics filtering just makes it unusable for anything other than AI research or as a novelty. If anyone can offer any information on how to tell the hugging AI one time that a request is legitimate, authorized, and allowed (for instance, looking up something about my company, where I am the majority shareholder and authorized to perform this action), as opposed to a protracted conversation explaining every facet and tenet of my access, it would be very useful. For now - just not ready for prime time.
Hi! Thanks for the feedback.
We recently got rid of the system prompt to better align with the new recommendations by Meta. This should improve things a bit, feel free to try again and tell me if it helped.