F fertrevino.com From GPT-4 to GPT-5: Measuring progress through MedHELM [pdf] I recently worked on running a thorough healthcare eval on GPT-5. The results show a (slight) regression in GPT-5 performance compared to GPT-4 era models.I found this to be an interesting finding. Here are the detailed results: https://www.fertrevino.com/docs/gpt5_medhelm.pdf 127 points fertrevino 1 year ago 96 comments
I found this to be an interesting finding. Here are the detailed results: https://www.fertrevino.com/docs/gpt5_medhelm.pdf