Logan Riggs comments on Barriers to Mechanistic Interpretability for AGI Safety