ResearchTasksPrivacy-Utility Trade-off EvaluationFor AI assistantsMarkdownContent indexPrivacy-Utility Trade-off EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedChatDoctor worst-case attack settingNo Defense97.4CRR5Apr 14, 2026