ESSENTIAL READING · PAPER 32
Enhancing LLM-Based Feedback: Insights from Intelligent Tutoring Systems and the Learning Sciences
Stamper, J., Xiao, R., & Hou, X. (2024). Enhancing LLM-Based Feedback: Insights from Intelligent Tutoring Systems and the Learning Sciences. arXiv preprint arXiv:2405.04645.
FOUND
Argues LLM feedback design should be driven by learning-science theory, not prompt tricks.
MEANS
AI feedback quality comes from theory, not prompt engineering â a lesson for every institution fine-tuning tutors.
SI DATA
No direct SI dataset counterpart yet — flagged for future audit expansion.