reviews. 2 Reviewer 1 comments that the experiments could be on larger problems - we agree that this is fair, and this is certainly 3 of interest going forward

We thank all our reviewers for their time and for their kind words. Reviewer 2 comments that some details are unclear. These details were put in the appendix for space; we will add them to the main paper. As you suggest, the model implements a feedforward step analogous to an RNN. On the number of function evaluations (NFE), leaving this out was an oversight.