The reviewers appreciated the primary focus of the paper on extending hyperbolic embeddings to multi-relational graphs, and found the proposed approach to be interesting and the results to be fairly strong (on WNET). The reviewers raised a number of concerns that were addressed in the authors' response, however, the additional content, such as the ablation studies and the results on the additional datasets, should be included in the main paper. One remaining concern was the performance when R and r are reversed, which seemed inconsistent with the proposed model (please clarify the results in the revision).