I just made up a rule that I want ChatGPT to follow every time I use it and I don’t want to have to put it in every time. I would like ChatGPT to look up my rules when we start a chat. I do see a strong probability that the rules could conflict with each other, thus leading to illogical replies.
My rule is “Please present labels/names of entities, or URL’s that exist first rather than making them up and presenting them as default when you answer my questions”
Note: I think this rule could be programmed into ChatGPT as a tweak and I would be fine with that as a solution to my immediate issue. It seems like a rule like this is fundamental: Look for real stuff first and lean towards that before making something up. My scope is restricted to labels or names, not to the conceptual level.
ChatGPT tends to give me answers that contain false names as a default even when it has the right information. I get irritated when I ask for information, get false labels as a result which I have to verify, then I take another step to ask for the real information, which it then provides. That really brings out the Karen in me. If this basic rule isn’t programmed into ChatGPT, then I want a way for ChatGPT to look it up. ChatGPT’s excuse that it doesn’t have access to all the available information doesn’t cover this scenario. Talk about a hit dog howl excuse.