Fine-tuning ChatGPT for automatic scoring

This study highlights the potential of fine-tuned ChatGPT (GPT-3.5) for automatically scoring student written constructed responses using example assessment tasks in science education. The application of ChatGPT in research and academic fields has greatly enhanced productivity and efficiency. Recent studies on ChatGPT based on OpenAI's generative model GPT-3.5 proved its superiority in predicting the natural language with high accuracy and human-like responses. GPT-3.5 has been trained over enormous online language materials such as journals and Wikipedia; however, direct usage of pre-trained GPT-3.5 is insufficient for automatic scoring as students do not utilize the same language as journals or Wikipedia,…

• NS
  National Science Foundation

  Award: 2101104
• IO
  Institute of Education Sciences

  Award: R305A160219