Observed Preferences
Benchmarks
TaskDatasetSOTA resultTrendResults
Observed Preferences Full Fit
0.092Modelling Loss
8
Observed Preferences No Distractors Fit Distractors Loss
0.154Modelling Loss
6
Observed Preferences Single-Stage Fit Two-Stage Loss
0.185Modelling Loss
6
Observed Preferences K-Fold Cross-Validation
0.153Modelling Loss
6