Goto

Collaborating Authors

 reviewer



fe4b8556000d0f0cae99daa5c5c5a410-AuthorFeedback.pdf

Neural Information Processing Systems

ThismakesROARmorereliable.4 Reviewer 1 (R1) re: portrayal of human studies: R1 correctly points out our portrayal of human stud-5 ies requires more nuance. We would be glad to correct this and will update the manuscript accordingly.6 As the reviewer assumed correctly, the gap between estimators is far larger than the variance.10 But as the reviewer points out, sometimes the curve itself provides additional information.12 This44 minimum deletion area is identified by perturbing and evaluating the model output without retraining.