Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Commonsense Reasoning on HellaSwag (val)

95.3Accuracy

GPT-4

28.7446.0263.380.58Mar 15, 2023Sep 9, 2023Mar 5, 2024Aug 30, 2024Feb 24, 2025Aug 21, 2025Feb 16, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2023.03
95.3
2026.02
91
2026.02
90.2
2026.02
90
2023.03
85.6
2023.03
85.5
2023.03
84.2
2026.02
82.4
2026.02
81.8
2026.02
77.6
2024.01
54.8
2024.07
50.7
2024.01
50.5
2024.01
49.1
2024.07
46.8
2024.01
46.2
2024.07
46
2024.07
45
2024.01
38.8
2024.07
38.8
2024.07
33.8
2024.07
33.4
2024.07
33.3
2024.07
33.2
2024.07
31.3