List your company

Language Modeling on OpenWebText (dev)

18.32Perplexity (Dev)GELU
18.288418.501718.71518.9283Jun 30, 2026

Evaluation results

MethodLinks
GELUEpoch=2, Number of layers=12L, Parameter-matched=true2026.0618.32
NC-FFNEpoch=2, Number of layers=12L, Parameter-matched=true2026.0618.36
GELUEpoch=1, Number of layers=12L, Parameter-matched=true2026.0619
NC-FFNEpoch=1, Number of layers=12L, Parameter-matched=true2026.0619.11