A camp in Model Interpretability that says we should reverse engineers model completely.

  • Understand every layer and parameter