After the rebuttal there is consensus among reviewers that the paper should be accepted. The authors provided very convincing feedback in their rebuttal response, including experiments with new baselines, and are strongly encouraged to supplement the paper with these (as we're sure they would)! The reviews for this paper were very detailed, and there is not much that a meta-summary could add. On a technical note, the authors could supplement their explanation somewhat; in that readers could "see that the density p(o) of the orderings can be taken out of the KL terms, but it is not clear why the dependence of x_{o_i} on x_{o_1}, ...,x_{o_{i-1}} seems to not matter in the KL-calculation", to quote verbatim.