List your company

Language Modeling on Clustered Pre-training Data Cluster 3

16.903Perplexity (PPL)Split MoE
16.8999216.9207116.941516.96229Mar 19, 2026

Evaluation results

MethodLinks
Split MoEModel Variant=Split MoE2026.0316.903
BaseModel Variant=Base2026.0316.98