WinoGrande
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Winogrande
80.71Accuracy
24
Winogrande
73.88Winogrande Score
22
Winogrande
77.74Score
22
WinoGrande 0-shot (test)
73.95Accuracy
21
WinoGrande (WG) (val)
78.3Accuracy
21
Winogrande (test)
1.027MAE
20
Winogrande
0.862Accuracy
19
Winogrande
73.2Accuracy (Pre-Attack)
19
Winogrande
74Winogrande Accuracy
18
Winogrande
69.06Accuracy
17
Winogrande
88.8Accuracy
16
WinoGrande (dev)
79.2Accuracy
16
WinoGrande S In-Distribution Llama-3.1-8B (train test)
78.09Accuracy
15
Winogrande (WinoG) (test)
63.8Accuracy
13
Winogrande XL
60.5Accuracy
13
Winogrande
0.626Kendall Correlation
13
Winogrande
2.14Accuracy Improvement
12
Winogrande
8ECE (15-bin)
11
WinoGrande
64.01Character-level Accuracy
11
Winogrande (WG) 1.3k (test)
1MAE
11
Winogrande
0.8NLL
10
Winogrande Estonian
64.4Accuracy
10
Winogrande
3.476LIS
10
Winogrande (test)
52.48AuROC
10
Winogrande Medium (In-Distribution)
83.99Accuracy
9