Context-rich mathematics item generation using generative language models.

• In Event: GSIC Poster Session II

Abstract

This project aims to investigate the viability of using a large language model to automatically generate appropriate contexts based on the target construct and numerical data. We fine-tuned a pre-trained language model to generate mathematics items. Results show that the fine-tuned model can achieve accuracy and control in automated item generation.

Author

• Olushola Olufemi Soyoye