Abstract: The evaluation of student code submissions is difficult due to the wide variety of syntax and logic applied to solve a similar problem. This research exploits transformer models, CodeBERT ...