Has anyone been able to come up with a way to keep track of GPT-4 performance over time? I'm told that the API is explicit about changes to models and that the Chat interface is not.
I made up an example here to illustrate, but it's just very annoying because sometimes it puts at the beginning, slowing down my interaction, and it now refuses to obey my prompts to leave caveats out.
https://chat.openai.com/share/1f39af02-331d-4901-970f-2f4b0e...
looking at your sample and using character count as a rough proxy for tokens, (465/(1581-465))*100 means they added ~42% token count cost to your response explicitly adding caveats which you dont want. fun!