Informal mathematical proofs produced directly by LLMs are notoriously prone to hallucination. Apparent elegant argumentations from AI frequently demand substantial manual verification work from human ...