Tony ChanWhy measuring “faithfulness” matters for AI model interpretationsAn interpretation is NEVER a good one if it does NOT reflect the behaviour of the model.Mar 4, 20221Mar 4, 20221