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

Common Sense Reasoning on HellaSwag (test)

83.9Accuracy

Llama-3.3-70B-Base

33.56446.63259.772.768Dec 18, 2025Dec 26, 2025Jan 3, 2026Jan 11, 2026Jan 19, 2026Jan 27, 2026Feb 5, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
83.9
2026.01
83.43
2026.01
81.44
2026.01
79.31
2026.02
79.3
2025.12
78.5
2026.02
78.2
2026.02
77.7
2026.02
77.4
2026.02
77.2
2026.02
76.5
2026.02
76.2
2026.01
75.72
2026.02
75.7
2026.01
74.78
2025.12
74.5
2026.01
74.13
2026.01
74.12
2026.01
73.75
2026.02
73.6
2026.02
73.2
2026.02
73.2
2025.12
72.1
2025.12
71.8
2025.12
70.2
2025.12
69.5
2026.02
69.2
2026.01
68.4
2025.12
66.5
2025.12
65.2
2026.01
64.58
2025.12
64.1
2025.12
63.5
2025.12
60.2
2026.01
59.49
2026.01
58.63
2025.12
55.8
2026.01
53.66
2026.01
52.77
2026.02
49
2026.02
47.7
2026.02
47
2026.02
47
2026.02
40.2
2026.02
35.5