ResearchDatasetsHelpsteer2FollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsAttribute-controlled Text GenerationHelpSteer2 Negative Representative Target Score (test)0.987Diversity12Instruction following evaluationHelpSteer2 (val)71.8Quality9Instruction FollowingHelpsteer2 Trivial78.22Accuracy8Self-evaluationHelpSteer2 (val)87.8Top-5 Accuracy3
Attribute-controlled Text GenerationHelpSteer2 Negative Representative Target Score (test)0.987Diversity12