ResearchBenchmarksRanking fidelity evaluation on ImageNet L-inf RobustBench (test)For AI assistantsMarkdownContent indexRanking fidelity evaluation on ImageNet L-inf RobustBench (test)Follow1Spearman Correlation (ρ)GF-Score0.7920.8460.90.954Apr 14, 2026Evaluation resultsMethodMethodLinksSpearman Correlation (ρ)GF-ScoreCalibration status=Cal...Calibration status=Calibrated2026.041GF-ScoreCalibration status=Unc...Calibration status=Uncalibrated2026.040.9OriginalCalibration status=Unc...Calibration status=Uncalibrated2026.040.8