Comment by ViscountPenguin

Fine-tuning language models for calibrated probability predictions has been a thing since the Bert era though, that's really nothing special.