LLMs offer tremendous opportunities for pedagogical agents to help students construct knowledge and develop problem-solving skills, yet many of these agents operate on a "one-size-fits-all" basis, limiting their ability to personalize support. To address this, we introduce Evidence-Decision-Feedback (EDF), a theoretical framework for adaptive scaffolding with LLM agents. EDF integrates elements of intelligent tutoring systems (ITS) and agentic behavior by organizing interactions around evidentiary inference, pedagogical decision-making, and adaptive feedback. We instantiate EDF through Copa, a Collaborative Peer Agent for STEM+C problem-solving. In an authentic high school classroom study, we show that EDF-guided interactions align feedback with students' demonstrated understanding and task mastery; promote scaffold fading; and support interpretable, evidence-grounded explanations without fostering overreliance.
@article{arxiv.2602.01415,
title = {Evidence-Decision-Feedback: Theory-Driven Adaptive Scaffolding for LLM Agents},
author = {Clayton Cohn and Siyuan Guo and Surya Rayala and Hanchen David Wang and Naveeduddin Mohammed and Umesh Timalsina and Shruti Jain and Angela Eeds and Menton Deweese and Pamela J. Osborn Popp and Rebekah Stanton and Shakeera Walker and Meiyi Ma and Gautam Biswas},
journal= {arXiv preprint arXiv:2602.01415},
year = {2026}
}
Comments
To appear as a full paper in the proceedings of the 27th International Conference on Artificial Intelligence in Education (AIED26)