Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Commonsense QA on HellaSwag
Loading...
79.1
Accuracy
Llama 3-8B
47.848
55.9615
64.075
72.1885
Oct 16, 2024
Jan 25, 2025
May 7, 2025
Aug 16, 2025
Nov 26, 2025
Mar 7, 2026
Jun 17, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Llama 3-8B
Paradigm=AR, Training...
2026.06
79.1
Falcon-7B
Paradigm=AR, Training...
2026.06
76.3
Llama 2-7B
Paradigm=AR, Training...
2026.06
76.2
OLMo-7B
Paradigm=AR, Training...
2026.06
75.6
LLaDA-8B
Paradigm=Masked Diffus...
2026.06
70.5
YOCO (Dense)
Model Scale=4B
2026.06
67.6
YOCO (CLSA)
Model Scale=4B, Attent...
2026.06
67.4
Transformer
Model Scale=4B
2026.06
66.7
Sumi-7B
Paradigm=Uniform Diffu...
2026.06
60
FP16
Backbone=LLaMA2-7B, Av...
2024.10
57.1
CMPQ
Backbone=LLaMA2-7B, Av...
2024.10
56.75
AWQ
Backbone=LLaMA2-7B, Av...
2024.10
56.4
GPTQ
Backbone=LLaMA2-7B, Av...
2024.10
55.99
QuIP
Backbone=LLaMA2-7B, Av...
2024.10
55.74
CMPQ
Backbone=LLaMA2-7B, Av...
2024.10
53.85
RTN
Backbone=LLaMA2-7B, Av...
2024.10
53.7
QuIP
Backbone=LLaMA2-7B, Av...
2024.10
53.41
RTN
Backbone=LLaMA2-7B, Av...
2024.10
51.68
GPTQ
Backbone=LLaMA2-7B, Av...
2024.10
49.05
Feedback
Search any
task
Search any
task