List your company

Language Modeling on Zyda 2

2.286LossCCGQA
2.28462.294052.30352.31295Oct 6, 2025

Evaluation results

MethodLinks
CCGQAParameters=1B, Training Tokens=300B, Architecture=Dense2025.102.286
MHAParameters=1B, Training Tokens=300B, Architecture=Dense2025.102.297
GQAParameters=1B, Training Tokens=300B, Architecture=Dense2025.102.297
CCAParameters=1B, Training Tokens=300B, Architecture=Dense2025.102.307
MLAParameters=1B, Training Tokens=300B, Architecture=Dense2025.102.321