Post-hoc vs. Intrinsic Explanations

Proxy Model

Simple and interpretable model trained post-hoc to mimic the behavior of a complex model, serving as an approximation to facilitate interpretation.

← Wstecz