ANLI
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ANLI MNLI + SNLI trained (dev)
50.4Accuracy (A1)
8
ANLI
93.8AUC
7
ANLI R3
67.1Accuracy
7
ANLI R1
78.5Accuracy
7
ANLI R3
34.77F1 Score
7
ANLI R2
32.62F1 Score
7
ANLI R1
34.93ANLI R1 F1
7
ANLI MNLI + SNLI + FEVER (test)
57Overall Score
7
ANLI MNLI + SNLI + ANLI + FEVER (dev)
0.738R1 Score
7
ANLI (test)
4.01GPT-4 Score
6
ANLI R3
62.4Clean Accuracy
4
ANLI Round 3 (R3)
44.43Macro F1
4
ANLI R3 (OOD)
32.12Accuracy
4
ANLI all rounds (test)
58.2Accuracy
4
ANLI R1
61.2ANLI R1 EM
3
ANLI R3
48.25Accuracy
3
ANLI-R2
40.5Accuracy
3
aNLI (dev)
77.94Accuracy
3
ANLI Out
12.27ECE
2
ANLI (R1+R2+R3 Average)
59.6ANLI Avg Accuracy
2
ANLI
61.5Accuracy
1