ResearchBenchmarksRanking fidelity evaluation on CIFAR-10 L2 RobustBench (test)For AI assistantsMarkdownContent indexRanking fidelity evaluation on CIFAR-10 L2 RobustBench (test)Follow0.897Spearman Rank CorrelationOriginal0.65260.716050.77950.84295Apr 14, 2026Evaluation resultsMethodMethodLinksSpearman Rank CorrelationOriginalCalibration status=Cal...Calibration status=Calibrated2026.040.897GF-ScoreCalibration status=Cal...Calibration status=Calibrated2026.040.871OriginalCalibration status=Unc...Calibration status=Uncalibrated2026.040.662GF-ScoreCalibration status=Unc...Calibration status=Uncalibrated2026.040.662