Teaching a Large Language Model Tutor to Withhold the Answer: A Supervisor Architecture and an Evidence-Driven Method for Tuning Socratic Behavior
A deployed tutoring system that enforces answer-withholding as a per-turn, machine-checkable contract, and a method for tuning that withholding against evidence is reported, offering the measure, diagnose, and fix loop as a reusable recipe for any LLM agent that must refuse a capability it has.