Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Language Modeling on (train)
Loading...
2.29
Perplexity
DeltaNet
2.2804
2.3452
2.41
2.4748
Mar 20, 2026
Perplexity
Loss
Updated 25d ago
Evaluation Results
Method
Method
Links
Perplexity
Loss
DeltaNet
Parameters=370M
2026.03
2.29
-
HADES
Parameters=218M
2026.03
2.31
-
Mamba2
Parameters=370M
2026.03
2.33
-
RetNet
Parameters=370M
2026.03
2.41
-
Linear Transformer
Parameters=370M
2026.03
2.49
-
Mamba1
Parameters=370M
2026.03
2.53
-
MoE-A0.5B-12B
Tokens=243B
2026.01
-
1.852
ConceptMoE-A0.5B-12B
Tokens=243B
2026.01
-
1.849
MoE-A1B-24B
Tokens=559B
2026.01
-
1.717
ConceptMoE-A1B-24B
Tokens=559B
2026.01
-
1.711
Feedback
Search any
task
Search any
task