ResearchDatasetsSafety Scaling TestbedFor AI assistantsMarkdownContent indexSafety Scaling TestbedFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsSafety Gate Capacity ScalingSafety Scaling Testbed N=100-10,00012.4Classifier Ceiling (B=0.01)3