Comment by johannesf

Have you done any fine-tuning or prompt-customization for the R-specific work? I've found the models worse on R when compared to Python, especially for more complex tasks. This looks cool, thanks for sharing!