ML Model Explanation

This is a prompt designed to help Claude explain how machine learning models make their predictions. You'd use it when you need to understand feature importance, generate SHAP values, create LIME explanations, or visualize what your model is actually looking at. It covers both global model behavior and local per-prediction breakdowns. The skill documentation is pretty thorough on explainability techniques, which is good because there's real nuance between something like partial dependence plots versus attention maps. Useful if you're debugging a model, need to satisfy compliance requirements, or just want to know why your neural network thinks that image is a cat.