List your company

Language Modeling on Qwen3 (val)

2.493Validation LossDana-MK4
2.46362.662052.86053.05895Feb 5, 2026

Evaluation results

MethodLinks
Dana-MK4Size=2.47B, kappa=0.852026.022.493—
ADanaSize=2.47B, kappa=0.852026.022.496—
Dana-Star-MK4Size=2.47B, kappa=0.852026.022.496—
Dana-MK4Size=2.47B, kappa=0.752026.022.501—
Dana-Star-MK4Size=2.47B, kappa=0.752026.022.512—
Dana-MK4Size=2.09B, kappa=0.852026.022.527—
Dana-Star-MK4Size=2.09B, kappa=0.852026.022.53—
ADanaSize=2.09B, kappa=0.852026.022.531—
AdamWSize=2.47B2026.022.536—
Dana-MK4Size=2.09B, kappa=0.752026.022.542—
Dana-Star-MK4Size=2.09B, kappa=0.752026.022.543—
AdemamixSize=2.09B2026.022.566—
AdamWSize=2.09B2026.022.572—
Dana-MK4Size=1.44B, kappa=0.852026.022.601—
ADanaSize=1.44B, kappa=0.852026.022.608—
Dana-Star-MK4Size=1.44B, kappa=0.852026.022.609—
Dana-MK4Size=1.44B, kappa=0.752026.022.618—
Dana-Star-MK4Size=1.44B, kappa=0.752026.022.62—
AdemamixSize=1.44B2026.022.641—
AdamWSize=1.44B2026.022.644—
Dana-MK4Size=947M, kappa=0.852026.022.692—
ADanaSize=947M, kappa=0.852026.022.696—
Dana-Star-MK4Size=947M, kappa=0.852026.022.699—
Dana-MK4Size=947M, kappa=0.752026.022.707—
Dana-Star-MK4Size=947M, kappa=0.752026.022.709—
AdemamixSize=947M2026.022.73—
AdamWSize=947M2026.022.74—
Dana-MK4Size=588M, kappa=0.852026.022.815—
ADanaSize=588M, kappa=0.852026.022.816—
Dana-Star-MK4Size=588M, kappa=0.852026.022.82—
Dana-MK4Size=588M, kappa=0.752026.022.834—
Dana-Star-MK4Size=588M, kappa=0.752026.022.839—
AdemamixSize=588M2026.022.868—
AdamWSize=588M2026.022.871—
ADanaSize=338M, kappa=0.852026.022.965—
Dana-MK4Size=338M, kappa=0.852026.022.965—
Dana-MK4Size=338M, kappa=0.752026.022.971—
Dana-Star-MK4Size=338M, kappa=0.852026.022.977—
Dana-Star-MK4Size=338M, kappa=0.752026.022.98—
AdemamixSize=338M2026.022.999—
AdamWSize=338M2026.023.017—
ADanaSize=176M, kappa=0.852026.023.182—
Dana-MK4Size=176M, kappa=0.852026.023.183—
Dana-MK4Size=176M, kappa=0.752026.023.185—
Dana-Star-MK4Size=176M, kappa=0.752026.023.199—
Dana-Star-MK4Size=176M, kappa=0.852026.023.202—
AdemamixSize=176M2026.023.214—
AdemamixSize=2.47B2026.023.226—
AdamWSize=176M2026.023.228—
CR-NetTraining steps=10K, Training tokens (B)=1.32025.09—25.99
CR-NetTraining steps=20K, Training tokens (B)=2.62025.09—21.73
CR-NetTraining steps=30K, Training tokens (B)=3.92025.09—19.32
CR-NetTraining steps=40K, Training tokens (B)=5.22025.09—18.12
Full-rankTraining steps=10K, Training tokens (B)=1.32025.09—24.31
Full-rankTraining steps=20K, Training tokens (B)=2.62025.09—20.53
Full-rankTraining steps=30K, Training tokens (B)=3.92025.09—18.84
Full-rankTraining steps=40K, Training tokens (B)=5.22025.09—17.85