ResearchTasksSelf-other distinctionSelf-other distinctionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedRobot evaluation setting (Held-out)Ours99.5Accuracy6Jun 12, 2026Robot + Human (Held-out)Ours99.69Accuracy6Jun 12, 2026