ESSENTIAL READING · PAPER 28
The promise and limits of LLMs in constructing proofs and hints for logic problems in intelligent tutoring systems
Tithi, S. D., Ramesh, A. K., DiMarco, C., Tian, X., et al. (2025). The promise and limits of LLMs in constructing proofs and hints for logic problems in intelligent tutoring systems. arXiv preprint arXiv:2505.04736.
FOUND
Mapped the accuracy ceiling â and the gaps â of LLM-generated hints for logic proofs.
MEANS
AI hints are usable, but the âwhyâ still needs teaching: hints without explanation build dependence, not understanding.
SI DATA
No direct SI dataset counterpart yet — flagged for future audit expansion.