Reproduce “TutorCaseBench: A Case-Grounded Benchmark for Evaluating LLM Tutors in Supporting Students with Challenging Behaviors” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yeon Su Park, Chanyoung Park, Tae Soo Kim, Seungju Kim, Hyunwoo Kim, Daho Jung, Jinsung Yoon, Juho Kim
TutorCaseBench evaluates whether language models can tailor tutoring responses to documented student challenges. The authors build synthesized text dialogues from educational case literature and assess open-ended replies against case-specific intervention criteria, alongside a strategy-selection task. Across 11 models, Claude Opus 4.7 achieves 55.8% rubric success, and selection scores are higher; these results measure responses to benchmark cases, not student learning.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.