Commonsense Reasoning on HellaSwag EL
52.58AccuracyLayerMoE-S1
Evaluation results
| Method | Links | |
|---|---|---|
| LayerMoE-S1Backbone=Llama-3.2-3B, Stage=Stage 1, #Experts=842026.03 | 52.58 | |
| LayerMoEBackbone=Llama-3.2-3B, Stage=Stage 2, #Experts=842026.03 | 52.41 | |
| NeuronMoE-S1Backbone=Llama-3.2-3B, Stage=Stage 1, #Experts=49 (-42%)2026.03 | 50.69 | |
| NeuronMoEBackbone=Llama-3.2-3B, Stage=Stage 2, #Experts=49 (-42%)2026.03 | 50.53 | |
| NeuronMoE-ENBackbone=Llama-3.2-3B, Stage=Stage 1, #Experts=37 (-56%)2026.03 | 49.03 | |
| Llama-3.2-3B DenseBackbone=Llama-3.2-3B, Stage=Dense, #Experts=-2026.03 | 43.38 | |
| LayerMoE-S1Backbone=Qwen-1.5-1.8B, Stage=Stage 1, #Experts=722026.03 | 35.21 | |
| NeuronMoE-S1Backbone=Qwen-1.5-1.8B, Stage=Stage 1, #Experts=36 (-50%)2026.03 | 31.26 |