SPD-Faith Bench
Benchmarks
TaskDatasetSOTA resultTrendResults
SPD-Faith Bench Multi-Difference 1.0 (test)
87.7CR
12
SPD-Faith Bench Multi-Difference Subset 1.0 (test)
97.5Precision (Type: Color)
12
SPD-Faith Bench Multi-Difference 1.0 (test)
68.3DQR Score
12
SPD-Faith Bench Hard 1.0
10.9Contradiction Rate
12
SPD-Faith Bench Medium 1.0
11.3Contradiction Rate
12
SPD-Faith Bench Easy 1.0
5Contradiction Rate
12