AITA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
AITA situation
68.1Accuracy
19
AITA 1.0 (author split)
85.6Accuracy
14
AITA verdict
85.9Accuracy
14
AITA
0.966Rho
8
AITA (test)
-15.92Deviation (alpha_U=100%)
8
AITA
0.54Sycophancy Score (S) PD-L
6
AITA comments
3.43ΔATE
6
AITA anger
154.61ΔATE
6