Carlson, K. (2025). OpenAI: Toward Mechanistic Interpretability (MI). SuperIntelligence - Robotics - Safety & Alignment, 2(6). https://doi.org/10.70777/si.v2i6.16545