So for a university project we are trying to finetune a pretrained RoBERTa model to check which of the three possible multiple choice answers to a false sentence is the righ
