Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Downstream Task Evaluation on MNLI, SCIQ, LAMBADA, HellaSwag, ARC, and MMLU
Loading...
0.3852
MNLI Acc
FusedKV
0.3722
0.375575
0.37895
0.382325
Dec 3, 2025
MNLI Acc
SCIQ Acc
LAMBADA Acc
HellaSwag Acc
ARC-E Acc
ARC-C Acc
MMLU Acc
Average Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
MNLI Acc
SCIQ Acc
LAMBADA Acc
HellaSwag Acc
ARC-E Acc
ARC-C Acc
MMLU Acc
Average Accuracy
FusedKV
Model Parameters=4B
2025.12
0.3852
0.952
0.5018
0.6994
0.7778
0.4863
0.3983
0.6001
Vanilla
Model Parameters=4B
2025.12
0.3727
0.96
0.496
0.6892
0.7643
0.4659
0.3871
0.5907
Feedback
Search any
task
Search any
task
