Evals 0/1

Evals

Interpretability

Explaining a model’s predictions with permutation importance, SHAP values, partial dependence and counterfactuals, and the traps in each, such as correlated features. Explanations are how a team checks that a model relies on the signals it should.

0 / 6 solved

Questions

0/6 solved
Question Tags Difficulty Success
1. What do SHAP values explain?
Machine LearningInterpretability
Medium —
2. What is a limitation of permutation importance?
Machine LearningInterpretability
Medium —
3. What is a counterfactual explanation?
Machine LearningInterpretability
Medium —
4. Why can permutation importance underrate correlated features?
Machine LearningInterpretability
Medium —
5. Zoom In, Zoom Out
interpretabilityglobal-explanations +1
Medium —
6. The Marginal Effect
interpretabilitypartial-dependence
Medium —