List your company

Language Modeling on Nemotron-CC 25M held-out tokens

9.58PerplexityMoE Transformer w/ dynamic conv. (low-rank)
9.152812.036414.9217.8036Jun 2, 2026

Evaluation results

MethodLinks
MoE Transformer w/ dynamic conv. (low-rank)Train Tokens=100B, Params=6.80B, Conv. Location=QKV2026.069.58
MoE Transformer w/ dynamic conv. (head-wise)Train Tokens=100B, Params=6.80B, Conv. Location=QKV2026.069.65
MoE Transformer w/ static conv.Train Tokens=100B, Params=6.77B, Conv. Location=QKV2026.069.74
MoE TransformerTrain Tokens=100B, Params=6.77B, Conv. Location=–2026.069.86
Transformer w/ dynamic conv. (low-rank)Train Tokens=100B, Params=1.88B, Conv. Location=all linear2026.0610.95
Transformer w/ dynamic conv. (low-rank)Train Tokens=100B, Params=1.88B, Conv. Location=QKV2026.0611.24
Transformer w/ dynamic conv. (head-wise)Train Tokens=100B, Params=1.88B, Conv. Location=QKV2026.0611.41
Transformer w/ static conv.Train Tokens=100B, Params=1.83B, Conv. Location=QKV2026.0611.5
Transformer w/ more params (wider FFN)Train Tokens=100B, Params=1.87B, Conv. Location=–2026.0611.67
TransformerTrain Tokens=100B, Params=1.82B, Conv. Location=–2026.0611.71
Transformer w/ dynamic conv. (low-rank)Train Tokens=15B, Params=319.0M, Conv. Location=all linear2026.0617.42
Gated DeltaNet w/ dynamic conv. (low-rank)Train Tokens=15B, Params=309.5M, Conv. Location=QKV2026.0617.95
Transformer w/ dynamic conv. (low-rank)Train Tokens=15B, Params=311.8M, Conv. Location=QKV2026.0618.01
Gated DeltaNet w/ dynamic conv. (head-wise)Train Tokens=15B, Params=309.6M, Conv. Location=QKV2026.0618.03
Transformer w/ dynamic conv. (head-wise)Train Tokens=15B, Params=311.7M, Conv. Location=QKV2026.0618.22
Transformer w/ static conv.Train Tokens=15B, Params=305.4M, Conv. Location=QKV2026.0618.66
Mamba-2 w/ dynamic conv. (head-wise)Train Tokens=15B, Params=309.8M, Conv. Location=QKV2026.0618.69
Mamba-2 w/ dynamic conv. (low-rank)Train Tokens=15B, Params=309.8M, Conv. Location=QKV2026.0618.72
Gated DeltaNet w/ static conv.Train Tokens=15B, Params=305.4M, Conv. Location=QKV2026.0618.75
Gated DeltaNet (w/o conv.)Train Tokens=15B, Params=305.2M, Conv. Location=–2026.0618.93
Transformer w/ more params (wider FFN)Train Tokens=15B, Params=311.5M, Conv. Location=–2026.0618.99
TransformerTrain Tokens=15B, Params=305.2M, Conv. Location=–2026.0619.12
Mamba-2 w/ static conv.Train Tokens=15B, Params=306.4M, Conv. Location=QKV2026.0619.3
Mamba-2 (w/o conv.)Train Tokens=15B, Params=306.2M, Conv. Location=–2026.0620.26