Best Practices for Custom GPTs Configurations

Hello everyone, do you think it’s a good practice to include examples of how to perform tasks in the configuration of a custom GPTs model?


I have found that using principles helps. For instance I built a model to evaluate writing. It has a list of good principles to follow. I instruct the model to check the text against these principles and recommend changes based on them.

I’ve done the same thing with Coding GPTs. I also instruct it to create new questions. What would a python programmer want to know about what’s contained in the text? Generate extra questions, etc.