for one part of my master’s thesis, I would like to fine-tune GPT4 turbo and benchmark it on the MultimedQA dataset.
My research will be based on these 2 papers here:
Capabilities of GPT-4 on Medical Challenge Problems
Large language models encode clinical knowledge
I have already tried to request access to fine-tuning GPT4 here, but I am not invited.
Would it be possible to get access if I can prove that the GPT4 fine-tuning would only be used for academic purposes at the University of Oxford?
I would obviously like to publish my results and I am confident that a fine-tuned GPT4 turbo could beat the previous SOTA on the MultimedQA.
Edit: had to delete all links…