List your company

Multilingual Understanding on M-MMLU

1.235Normalized PerformanceTask Arithmetic
0.796120.910061.0241.13794Jun 15, 2026

Evaluation results

MethodLinks
Task ArithmeticCategory=Conventional Method (+Many-shot), Backbone=Llama-3.2-3B2026.061.235
TA + Subspace BoostingCategory=State-of-the-Art (+Many-shot), Backbone=Llama-3.2-3B2026.061.229
Iso-CCategory=State-of-the-Art (+Many-shot), Backbone=Llama-3.2-3B2026.061.018
Iso-CTSCategory=State-of-the-Art (+Many-shot), Backbone=Llama-3.2-3B2026.061.012
FedMergeCategory=Iterative / FL, Backbone=Llama-3.2-3B2026.061.012
TSV-MCategory=State-of-the-Art (+Many-shot), Backbone=Llama-3.2-3B2026.060.964
TULU 3Category=Data-Mixing, Backbone=Llama-3.2-3B2026.060.934
q-FedAvgCategory=Iterative / FL, Backbone=Llama-3.2-3B2026.060.916
METISCategory=Ours, Backbone=Llama-3.2-3B2026.060.91
DARECategory=Conventional Method (+Many-shot), Backbone=Llama-3.2-3B2026.060.904
ColD FusionCategory=Iterative / FL, Backbone=Llama-3.2-3B2026.060.837
ConsensusTACategory=Conventional Method (+Many-shot), Backbone=Llama-3.2-3B2026.060.825
TIESCategory=Conventional Method (+Many-shot), Backbone=Llama-3.2-3B2026.060.813