Pabulib
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Pabulib large (test)
1Omega Prime (sat^cost)
30
Pabulib small (test)
1omega'[sat^cost]
30
Pabulib OOD (test)
1W'[sat_card]
15
Pabulib ID (test)
100W'[sat_card]
15
Pabulib-large (test)
1omega'[sat^card]
15
Pabulib small (test)
1Omega'[sat^card]
15
Pabulib approval ballots v1 (OOD)
1w'[sat_cost]
15
Pabulib cardinal ballots
1Omega Prime (ω')
8
Pabulib 20 PB
85Representational Accuracy
2