Okay, I didn't think of that on first thought, but I guess it's best to take the conservative approach, and only allow what's understood. It's almost like the p

Okay, I didn't think of that on first thought, but I guess it's best to take the conservative approach, and only allow what's understood. It's almost like the principle of least privilege for the response, there's probably a better name for it. It could also be done on the request side, and I have seen some examples. I guess, prompt leaking at the end of the day isn't that terrible... I don't know, just brainstorming out loud. How unique and valuable are prompts going to be? Probably less valuable as models progress.