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

Accuracy on Winogrande (Common-sense Reasoning)

85.64Accuracy

ST-D

49.926459.198268.4777.7418Nov 6, 2025Dec 15, 2025Jan 24, 2026Mar 4, 2026Apr 13, 2026May 22, 2026Jul 1, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.06
85.64
2026.06
85.16
2026.06
84.93
2026.06
84.61
2026.06
84.53
2026.06
78.22
2026.06
77.82
2026.06
77.74
2025.11
77.2
2026.06
76.48
2026.06
76.48
2026.07
76.48
2025.11
75.8
2025.11
75.5
2025.11
75.4
2026.06
75.14
2025.11
75.1
2025.11
75
2025.11
75
2025.11
74.5
2026.06
74.03
2026.07
73.16
2025.11
73
2025.11
73
2025.11
72.7
2026.06
72.45
2025.11
72.1
2025.11
72
2025.11
70.6
2025.11
70.5
2025.11
70.4
2025.11
70.4
2025.11
70.3
2025.11
70.1
2025.11
70.1
2025.11
70
2025.11
69.6
2025.11
69.4
2026.07
69.3
2025.11
69.2
2025.11
69.1
2025.11
69
2026.07
68.98
2025.11
68.9
2026.06
68.35
2025.11
67.8
2026.06
67.32
2025.11
66.8
2026.06
66.38
2025.11
66.1
2025.11
66
2025.11
65.7
2026.06
65.51
2026.06
65.43
2025.11
65
2025.11
65
2026.07
63.54
2025.11
63.2
2025.11
63
2025.11
60.3
2025.11
60.2
2025.11
56.1
2025.11
55.3
2025.11
54
2026.06
53.59
2026.07
53.12
2026.06
51.3