robust hearth
Hey everyone! I've been exploring the new GPTs feature and have a few questions about how the GPT configuration works. Specifically:
Does the entire GPT configuration reside only in the blurb listed under 'configure', 'instructions' and the other fields for documents etc. or are there other aspects to it?
I've been giving constant feedback to some GPTs using the assistant. Does this feedback change the GPT's configuration beyond what's described in the 'instructions' blurb? Does anything happen on the back end?
Is it more effective to manually create a GPT using feedback and the built-in assistant, or should you edit the blurb yourself for better consistency and results?
Do the GPTs have any hidden, server-side configuration that updates through interactions with the GPT building assistant?
Thanks!