List your company

Language Modeling on Clustered Pre-training Data Cluster 14

20.656Perplexity (PPL)Split MoE
20.6508820.6854420.7220.75456Mar 19, 2026

Evaluation results

MethodLinks
Split MoEModel Variant=Split MoE2026.0320.656
BaseModel Variant=Base2026.0320.784