This study explores leveraging GPT models, specifically GPT-3.5 and GPT-4, to create automated feedback in tutoring by employing sequence labeling techniques like prompt engineering and fine-tuning. Focusing on identifying praise components in tutor responses, the research outlines how carefully crafted prompts can enhance the models' understanding and generation of feedback. Results indicate that both methods significantly improve feedback quality, demonstrating the potential of large language models in educational contexts and tutor training.
Prompt engineering involves designing input prompts to guide GPT models in generating desired outputs; it optimizes the model's ability to identify praise in tutor responses.
The study highlights how using fine-tuning and prompt engineering can significantly enhance large language models like GPT-3.5 in providing effective feedback through sequence labeling.
Collection
[
|
...
]