List your company

Language Modeling on Public Pretraining Dataset OOD

1.606LossNexus
1.595721.665111.73451.80389Apr 10, 2026

Evaluation results

MethodLinks
NexusModel scale=3B2026.041.606
AdamModel scale=3B2026.041.623
NexusModel scale=1B2026.041.835
AdamModel scale=1B2026.041.863