List your company

Language Modeling on ClimbMix FineWeb-Edu 1M-token (val)

30.12PPLIDN Reg.
28.070841.902955.73569.5671May 18, 2026

Evaluation results

MethodLinks
IDN Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=35.31, Regularization Gain=-5.0%, Config=12xF1-h0, K=82026.0530.12-14.70.5779.999.39595.5
DiagN Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=35.41, Regularization Gain=-4.7%, Config=12xF1-h0, K=82026.0531.39-11.30.5779.29990.787.7
No Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=37.16, Config=12xF1-h0, K=82026.0531.45-15.40.5980.398.994.392.2
DiagN Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=35.41, Regularization Gain=-4.7%, Config=4xF1-fwd, K=22026.0531.52-110.9287.798.994.868.8
No Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=37.16, Config=4xF1-fwd, K=22026.0532.27-13.20.9487.299.295.578.3
IDN Reg.Model Architecture Variant=Nanochat-3B standard, Sequential PPL Baseline=35.31, Regularization Gain=-5.0%, Config=4xF1-fwd, K=22026.0532.96-6.70.9990.599.896.991
IDN Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=53.25, Regularization Gain=-23.4%, Config=2xF6-fwd, K=12026.0544-17.41.3768.998.893.598.8
No Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=69.54, Config=16xF1-h0, K=82026.0547.25-32.10.7850.997.788.972.3
DiagN Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=63.08, Regularization Gain=-9.3%, Config=12xF1-h0, K=82026.0551.42-18.50.8864.998.494.390.8
IDN Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=53.25, Regularization Gain=-23.4%, Config=12xF2-h0, K=22026.0553.680.82.3753.297.987.935.3
No Reg.Model Architecture Variant=Nanochat-0.5B-mHC, Sequential PPL Baseline=73.24, Config=4xF2-fwd, K=22026.0561.34-16.21.1366.79995.589.2
No Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=69.54, Config=12xF1-h0, K=42026.0562.01-10.81.1162.698.295.261.4
DiagN Reg.Model Architecture Variant=Nanochat-0.5B standard, Sequential PPL Baseline=63.08, Regularization Gain=-9.3%, Config=4xF2-h0, K=22026.0563.40.51.176197.993.850.4
HCN Reg.Model Architecture Variant=Nanochat-0.5B-mHC, Sequential PPL Baseline=67.23, Regularization Gain=-8.2%, Config=8xF1-h0, K=12026.0565.91-20.9791.199.898.882.8
HCN Reg.Model Architecture Variant=Nanochat-0.5B-mHC, Sequential PPL Baseline=67.23, Regularization Gain=-8.2%, Config=20xF1-h0, K=42026.0566.56-11.2291.199.398.680.1
No Reg.Model Architecture Variant=Nanochat-0.5B-mHC, Sequential PPL Baseline=73.24, Config=4xF3-fwd, K=22026.0569.42-5.21.3156.19893.771.1
IDN Reg.Model Architecture Variant=Nanochat-0.5B w/o x0ve, Sequential PPL Baseline=79.96, Regularization Gain=-5.6%, Config=4xF4-h0, K=42026.0572.71-9.11.247499.695.295.9
IDN Reg.Model Architecture Variant=Nanochat-0.5B w/o x0ve, Sequential PPL Baseline=79.96, Regularization Gain=-5.6%, Config=4xF6-h0, K=22026.0575.09-6.12.3273.699.695.284.5
No Reg.Model Architecture Variant=Nanochat-0.5B w/o x0ve, Sequential PPL Baseline=84.74, Config=6xF2-fwd, K=42026.0578.54-7.31.0590.799.898.697.9
No Reg.Model Architecture Variant=Nanochat-0.5B w/o x0ve, Sequential PPL Baseline=84.74, Config=8xF2-fwd, K=22026.0581.35-41.3976.199.296.555.3