List your company

Common-Sense Reasoning on Common-Sense Reasoning Tasks

16Wiki PPLMamba2-H
15.488818.939422.3925.8406Oct 8, 2025

Evaluation results

MethodLinks
Mamba2-HParameters=1.3B, Training Tokens=100B, Architecture Type=Hybrid2025.101612.7757.7428.2471.9352.7348.4655.4952.43
GDN-HParameters=1.3B, Training Tokens=100B, Architecture Type=Hybrid2025.1016.0211.6858.1227.9970.5752.5249.7856.2752.54
NHAParameters=1.3B, Training Tokens=100B, Architecture Type=Hybrid2025.1016.1612.5858.7128.572.0352.0849.1956.8352.89
GSA-HParameters=1.3B, Training Tokens=100B, Architecture Type=Hybrid2025.1016.2223.0957.7928.0771.7151.741.9253.3550.76
GSAParameters=1.3B, Training Tokens=100B, Architecture Type=Base2025.1016.6912.6258.3328.3372.2550.9847.4353.4351.79
GDNParameters=1.3B, Training Tokens=100B, Architecture Type=Base2025.1017.1411.3556.8227.3971.7649.7749.151.7851.1
GLA-HParameters=1.3B, Training Tokens=100B, Architecture Type=Hybrid2025.1017.5815.7256.7327.5670.448.7146.7154.4650.76
Trans++Parameters=1.3B, Training Tokens=100B, Architecture Type=Base2025.1017.6115.8655.0128.0770.0849.2145.656.2750.71
GLAParameters=1.3B, Training Tokens=100B, Architecture Type=Base2025.1017.6115.3855.1827.5669.8648.8946.0553.9150.24
GDN-HParameters=340M, Training Tokens=15B, Architecture Type=Hybrid2025.1025.6751.7945.7523.2165.5635.829.351.3841.83
Mamba2-HParameters=340M, Training Tokens=15B, Architecture Type=Hybrid2025.1025.8143.3548.2324.4965.2336.3130.8648.6242.29
NHAParameters=340M, Training Tokens=15B, Architecture Type=Hybrid2025.1025.9738.3847.6924.3265.6736.0332.7252.0943.09
GSA-HParameters=340M, Training Tokens=15B, Architecture Type=Hybrid2025.1026.1946.0545.5423.7265.2335.1930.9150.9941.93
GDNParameters=340M, Training Tokens=15B, Architecture Type=Base2025.1026.4731.6946.0423.5566.0535.1834.3550.8342.67
Trans++Parameters=340M, Training Tokens=15B, Architecture Type=Base2025.1026.8842.1544.9125.9464.3134.9532.8451.0742.34
GLA-HParameters=340M, Training Tokens=15B, Architecture Type=Hybrid2025.1027.9440.1245.224.3263.9334.4231.8349.9641.61
GSAParameters=340M, Training Tokens=15B, Architecture Type=Base2025.1028.1742.5745.524.2364.853530.7850.4341.8
GLAParameters=340M, Training Tokens=15B, Architecture Type=Base2025.1028.783944.5322.2763.9334.8432.2751.3841.54