List your company
ResearchFormerly WizWand

General Language Understanding on Average

72.93Average AccuracyEMoE
33.805243.962654.1264.2774Sep 26, 2025
Updated 5mo ago

Evaluation results

MethodLinks
EMoEBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=12 + 22025.0972.930.55
EMoEBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=6 + 22025.0971.750.26
Top-kBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=6 + 22025.0971.60.23
Top-kBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=12 + 22025.0970.960.41
EMoEBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=3 + 22025.0970.450.12
Top-kBase Model=ERNIE-4.5-21B-A3B, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=3 + 22025.0969.640.15
EMoEBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=12 + 22025.0954.190.33
EMoEBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=6 + 22025.0952.780.18
Top-kBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=6 + 22025.0951.070.01
Top-kBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=12 + 22025.0950.940.16
EMoEBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=EMoE, Inference Budget (k')=3 + 22025.0950.130.64
EMoEBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=EMoE, Inference Budget (k')=62025.0949.50.65
EMoEBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=EMoE, Inference Budget (k')=42025.09490.7
EMoEBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=EMoE, Inference Budget (k')=22025.0948.220.62
Top-kBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=Top-k, Inference Budget (k')=42025.0948.050.36
Top-kBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=Top-k, Inference Budget (k')=22025.0947.880.57
Top-kBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=Top-k, Inference Budget (k')=62025.0947.570.52
Top-kBase Model=DeepSeek-V2-Lite, Trained Activated Experts=6+2, MoE Routing Method=Top-k, Inference Budget (k')=3 + 22025.0947.180.48
EMoEBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=EMoE, Inference Budget (k')=12025.0945.681.19
Top-kBase Model=LoRAMoE, Trained Activated Experts=2, MoE Routing Method=Top-k, Inference Budget (k')=12025.0945.480.91
EMoEBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=EMoE, Inference Budget (k')=162025.0943.130.74
EMoEBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=EMoE, Inference Budget (k')=82025.0941.970.59
Top-kBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=Top-k, Inference Budget (k')=82025.0941.350.55
Top-kBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=Top-k, Inference Budget (k')=162025.0941.090.75
EMoEBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=EMoE, Inference Budget (k')=42025.0937.810.72
Top-kBase Model=OLMoE-1B-7B-0924, Trained Activated Experts=8, MoE Routing Method=Top-k, Inference Budget (k')=42025.0935.311.06