Different result every time code is rerun

[katharinawoerzner]

Hi,
I get a different result every time I run the code:
gptcelltype(markers, model = 'gpt-4', topgenenumber = 10, tissuename="CD8+ T cells")
even if I don't change any of the parameters or any of the inputs, but simply by rerunning it. All clusters change around so the one that was cytotoxic CD8+ cells before is now regulatory CD8+ T cells, and the time after they are suddenly naive.

As far as I can understand, there should not be any randomization in the code?
Perhaps it is not strong enough to use on subset populations like CD8+ T cells?

Did anyone else experience the same issue?

[alohasiqi]

I had the same issue. I believe that's a feature of using the OpenAI GPT model, similar to the way you would interact with ChatGPT.