(1)
Li, L. Evaluating the Architectural Reasoning Capabilities of LLM Provers via the Obfuscated Natural Number Game. AAAI-SS 2026, 8, 588-591.