List your company

Commonsense QA on HellaSwag

79.1AccuracyLlama 3-8B
47.84855.961564.07572.1885Oct 16, 2024Jan 25, 2025May 7, 2025Aug 16, 2025Nov 26, 2025Mar 7, 2026Jun 17, 2026

Evaluation results

MethodLinks
Llama 3-8BParadigm=AR, Training Tokens=15T, Training Data=Not Released, Evaluation protocol=Reported by prior work, Shots=02026.0679.1
Falcon-7BParadigm=AR, Training Tokens=1.5T, Training Data=Partially Released, Evaluation protocol=Evaluated under our protocol, Shots=02026.0676.3
Llama 2-7BParadigm=AR, Training Tokens=2T, Training Data=Not Released, Evaluation protocol=Evaluated under our protocol, Shots=02026.0676.2
OLMo-7BParadigm=AR, Training Tokens=2.5T, Training Data=Fully Released, Evaluation protocol=Evaluated under our protocol, Shots=02026.0675.6
LLaDA-8BParadigm=Masked Diffusion, Training Tokens=2.3T, Training Data=Not Released, Evaluation protocol=Reported by prior work, Shots=02026.0670.5
YOCO (Dense)Model Scale=4B2026.0667.6
YOCO (CLSA)Model Scale=4B, Attention=Cross-Layer Sparse2026.0667.4
TransformerModel Scale=4B2026.0666.7
Sumi-7BParadigm=Uniform Diffusion, Training Tokens=1.5T, Training Data=Fully Released, Evaluation protocol=Evaluated under our protocol, Shots=02026.0660
FP16Backbone=LLaMA2-7B, Avg. Bit=16, Evaluation Protocol=zero-shot2024.1057.1
CMPQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=zero-shot2024.1056.75
AWQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=zero-shot2024.1056.4
GPTQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=zero-shot2024.1055.99
QuIPBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=zero-shot2024.1055.74
CMPQBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=zero-shot2024.1053.85
RTNBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=zero-shot2024.1053.7
QuIPBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=zero-shot2024.1053.41
RTNBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=zero-shot2024.1051.68
GPTQBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=zero-shot2024.1049.05