Gevindu Ganganath, Pasindu Bolonghege, Qianru Lyu, Pradeep Varakantham, Thivya Kandappu (2026) — arXiv:2606.22349 (cs.CL; cs.HC) — submitted to EMNLP 2026 📄 Full text (arXiv)
Ganganath et al. (2026) introduce CURIOBOT, a framework that operationalizes Berlyne's four collative variables (novelty, complexity, conflict, uncertainty) as adaptive linguistic interventions in conversational tutoring. Across 270 tutoring conversations spanning multiple LLM model families, domains, and topic complexity levels, curiosity-oriented interventions consistently increased exploratory learner behaviors, producing up to 2.4x more conversational turns under fixed time budgets. A learner-centered evaluation framework captured exploratory questioning, conversational agency, productive struggle, and observable curiosity. Critically, learner-side gains persisted even when tutor-side instructional quality remained unchanged, suggesting that curiosity functions as a partially independent interaction-level mechanism — not merely a byproduct of instruction quality. This work demonstrates that LLM-mediated dialogue can serve as a scalable experimental framework for studying how language shapes exploratory learning behavior, with direct implications for metacognition and self-regulated-learning research, intelligent-tutoring design, and scaffolding strategies.
Related Pages
- curiobot-llm-tutoring-exploratory-learning -- This page