Deep learning theory, Model interpretability, Causal reasoning and analysis, Unsupervised learning Circuit level understanding, Mechanistic Insights, Reverse Engineering,