Analyzing Japanese Student Tutor Reports with LLM

Abstract

This study analyzes paper-based reports submitted by Japanese student tutors (hereafter referred to as tutors) who supported international students at a Japanese university. We compare the experiences of tutors who supported short-term exchange students with those who supported regular students enrolled in full academic programs using a Large Language Model (LLM). Summaries generated by the LLM show that tutors supporting exchange students focused more on short-term daily life assistance and cultural exchange, while tutors for regular students provided more academic support. Sentiment analysis revealed that regular student support was perceived as more burdensome.