Reviewer # 1

To clarify the origins of ASVs, we will modify lines 146-7: "In the game theory literature, this axiom was first relaxed To clarify the notion of accuracy in the global Shapley sum rule, we will add: "The accuracy of randomly drawing from Sec 4.3) could occur at any point in the time series: Each time series We will add a sentence in the text to clarify this and hope this makes the application seem less odd. Regarding R1's concern about the inefficiency of ASVs for feature selection, we propose to reframe Sec 4.4 as Please also see lines 29-32 below in our response to R3. R3's largest concern is that our paper does not discuss the difference between our approach and [19], which appears We will make the following addition to the end of Sec 3.2: Note that this is quite distinct from other work [19], which considers the model's prediction process itself In contrast, ASVs incorporate causal structure present in the data itself." R3 finds ASVs' incorporation of causality to be mainly based on intuition. We will clarify this in our introduction to ASVs. R3 is correct that the ASVs of Sec 4.2 place gender and department choice out-of-causal ordering.