Leveraging Pre-Trained Large Language Models for Automatic Item Generation: A Scoping Review

• In Event: Automatic Item Generation

Abstract

This scoping review examined 60 research studies that utilized pre-trained language models (PLMs) and large language models (LLMs) for automatic item generation (AIG). We summarized popular models, their specific usage in AIG, model performance, and the characteristics of items generated, underscoring the significant potential of PLMs and LLMs in AIG.

Authors

• Bin Tan, University of Alberta

• Nour Armoush, Unviersity of Alberta

• Okan Bulut, University of Alberta

• Elisabetta Mazzullo

• Mark Gierl, UNIVERSITY OF ALBERTA