ResearchTasksHelpfulness alignmentHelpfulness alignmentFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedHHH alignmentAMBS98.6Win Rate (WR)44May 19, 2026Anthropic hh-rlhfSHF-CAS3.36Gold Reward14Feb 26, 2026