List your company

Language Modeling on Clustered Pre-training Data Cluster 15

18.294Perplexity (PPL)Split MoE
18.2901618.3160818.34218.36792Mar 19, 2026

Evaluation results

MethodLinks
Split MoEModel Variant=Split MoE2026.0318.294
BaseModel Variant=Base2026.0318.39