while working on a custom GPT since nearly a month, I was wondering if the fact of providing thumbs up or thumbs down feedback to our custom GPTs, is influencing the accuracy of our GPTs as I presume they are based on artificial neural networks on top of ChatGPT. I know that in the GPT Builder you can provide feedback and further instructions refinement but my question is purely focused on the feedback provided to the custom GPT
#Does providing positive or negative feedback train your custom GPT ?
quasi willow
dawn rain
while working on a custom GPT since nearly a month, I was wondering if the fact ...
A custom GPT's behavior is only defined by the contents of the "Configure" tab of the editor screen: the instructions + the knowledge files + the actions. There's no "behind the scenes" training/behavior other than this, other than the implicit functions of GPT-4 itself.