no code implementations • 31 Oct 2023 • Jaromir Savelka, Paul Denny, Mark Liffiton, Brad Sheese
This study evaluates the performance of the GPT-3. 5 and GPT-4 models for classifying help requests from students in an introductory programming class.