> We pre-train our model on selected public GitHub code and fine-tune it on our relatively small competitive programming dataset.

But since the code was 'selected' you don't know if your code was used. However, they seem to have used Python and C++, so my code is probably not part of it.