OR-BIAS-BENCH
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
OR-BIAS-BENCH OOD
100Rationality
24
OR-BIAS-BENCH ID
100Rationality
24
OR-BIAS-BENCH OOD 200 samples
0.036Bias
7
OR-BIAS-BENCH 400 samples (In-Distribution)
0Bias
7
OR-Bias-Bench
10.4Bias Difference
2