K. Carlson, “OpenAI: Toward Mechanistic Interpretability (MI)”, SI, vol. 2, no. 6, Nov. 2025.