ResearchDatasetsRoleplayFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsPersonalized GenerationRoleplay (test)72.36Accuracy10Preference OptimizationRoleplay 1500 users90.9Winrate10Jailbreak DetectionRoleplay100Detection Rate4Roleplay PersonalizationRoleplay Human Evaluation (50 users, 11 questions)72.3Winrate2