Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Natural Language Inference on WANLI

45.4Accuracy (WANLI)

DOPA*

31.495235.105138.71542.3249Apr 13, 2026
Updated 1d ago

Evaluation Results

MethodLinks
2026.04
45.4
2026.04
42.77
2026.04
42.7
2026.04
42.37
2026.04
41.93
2026.04
41.77
2026.04
41.6
2026.04
41.33
2026.04
41.3
2026.04
41.13
2026.04
41
2026.04
40.87
2026.04
40.8
2026.04
40.7
2026.04
40.57
2026.04
40.07
2026.04
39.73
2026.04
39.47
2026.04
38.97
2026.04
38.87
2026.04
38.83
2026.04
38.4
2026.04
37.7
2026.04
37.67
2026.04
37.2
2026.04
36.5
2026.04
36.43
2026.04
36.23
2026.04
34.93
2026.04
34.23
2026.04
34
2026.04
33.8
2026.04
33.73
2026.04
33.7
2026.04
33.57
2026.04
33.5
2026.04
33.43
2026.04
33.37
2026.04
33.37
2026.04
33.27
2026.04
32.93
2026.04
32.03