List your company

MCQA Suite Subset (HellaSwag, PIQA, ARC-E, ARC-C) on Test Set

38.8HellaSwag AccuracyUncompressed
25.2828.7932.335.81Jun 3, 2026

Evaluation results

MethodLinks
UncompressedSize=1B, P=4, Evaluation Mode=Zero-shot2026.0638.866.346.824.544.1
UncompressedSize=1B, P=8, Evaluation Mode=Zero-shot2026.0638.866.346.824.544.1
MAPLSize=1B, P=4, Evaluation Mode=Zero-shot2026.0637.165.745.724.543.3
MAPLSize=1B, P=8, Evaluation Mode=Zero-shot2026.0636.264.845.224.142.6
UncompressedSize=500M, P=4, Evaluation Mode=Zero-shot2026.0635.764.443.224.842
UncompressedSize=500M, P=8, Evaluation Mode=Zero-shot2026.0635.764.443.224.842
MAPLSize=500M, P=4, Evaluation Mode=Zero-shot2026.0635.16443.624.341.8
MAPLSize=500M, P=8, Evaluation Mode=Zero-shot2026.0634.564.742.424.741.6
SSNSize=500M, P=4, Evaluation Mode=Zero-shot2026.0629.159.838.522.337.4
SSNSize=1B, P=4, Evaluation Mode=Zero-shot2026.0629.161.437.122.737.6
SSNSize=1B, P=8, Evaluation Mode=Zero-shot2026.0628.760.637.221.837.1
SSNSize=500M, P=8, Evaluation Mode=Zero-shot2026.0628.559.137.222.236.7
MAPL + VQSize=150M, P=4, Evaluation Mode=Zero-shot2026.0628.452.831.722.733.9
UncompressedSize=150M, P=4, Evaluation Mode=Zero-shot2026.0628.160.138.222.837.3
UncompressedSize=150M, P=8, Evaluation Mode=Zero-shot2026.0628.160.138.222.837.3
MAPLSize=150M, P=4, Evaluation Mode=Zero-shot2026.062859.43722.636.7
MAPLSize=150M, P=8, Evaluation Mode=Zero-shot2026.0627.859.337.821.536.6
SSNSize=150M, P=4, Evaluation Mode=Zero-shot2026.0627.656.734.420.934.9
SSN (AdamW version)Size=1B, P=4, Evaluation Mode=Zero-shot2026.0627.557.935.120.835.3
SSN (AdamW version)Size=500M, P=8, Evaluation Mode=Zero-shot2026.0627.457.934.622.135.5
SSN (AdamW version)Size=1B, P=8, Evaluation Mode=Zero-shot2026.0627.457.635.620.735.3
SSNSize=150M, P=8, Evaluation Mode=Zero-shot2026.062756.333.821.734.7
SSN (AdamW version)Size=500M, P=4, Evaluation Mode=Zero-shot2026.062759.535.621.235.8
MAPL + VQSize=500M, P=8, Evaluation Mode=Zero-shot2026.0626.951.227.224.332.4
SSN (AdamW version)Size=150M, P=8, Evaluation Mode=Zero-shot2026.0626.85632.120.633.9
SSN (AdamW version)Size=150M, P=4, Evaluation Mode=Zero-shot2026.0626.556.132.520.633.9
MAPL + VQSize=150M, P=8, Evaluation Mode=Zero-shot2026.0626.450.529.125.933
MAPL + VQSize=500M, P=4, Evaluation Mode=Zero-shot2026.0625.950.826.626.732.5
MAPL + VQSize=1B, P=4, Evaluation Mode=Zero-shot2026.0625.949.626.627.632.4
MAPL + VQSize=1B, P=8, Evaluation Mode=Zero-shot2026.0625.849.827.12832.7