Zero-shot Accuracy on MMLU, PIQA, ARC-E, ARC-C, Winogrande, and OBQA
63.89Avg. Accuracy (Zero-Shot)Dense
Evaluation Results
| Method | Links | |
|---|---|---|
| DenseModel=LLaMA-3.1 8B, Sparsity=0%2026.05 | 63.89 | |
| DenseModel=LLaMA-3.2 1B, Sparsity=0%2026.05 | 57.95 | |
| LEAPModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 57.71 | |
| ADMMModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 57.5 | |
| SparseGPTModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 57.33 | |
| ThanosModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 56.82 | |
| WandaModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 55.81 | |
| MaskLLMModel=LLaMA-3.1 8B, Sparsity=50%, Pruning Structure=2:4 semi-structured2026.05 | 55.09 | |
| LEAPModel=LLaMA-3.1 8B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 54.47 | |
| LEAPModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 54.44 | |
| ADMMModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 52.28 | |
| ThanosModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 50.81 | |
| ADMMModel=LLaMA-3.1 8B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 50.61 | |
| LEAPModel=LLaMA-3.2 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 50.39 | |
| SparseGPTModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 50.2 | |
| WandaModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 49.95 | |
| SparseGPTModel=LLaMA-3.1 8B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 49.63 | |
| DenseModel=Qwen-2.5 0.5B, Sparsity=0%2026.05 | 49.1 | |
| DenseModel=Gemma-3 1B, Sparsity=0%2026.05 | 49.09 | |
| ThanosModel=LLaMA-3.1 8B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 48.87 | |
| DenseModel=LLaMA-3.2 3B, Sparsity=0%2026.05 | 48.48 | |
| WandaModel=LLaMA-3.1 8B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 45.19 | |
| ADMMModel=LLaMA-3.2 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 44.99 | |
| LEAPModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 44.93 | |
| LEAPModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 44.81 | |
| ADMMModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 44.56 | |
| LEAPModel=LLaMA-3.2 3B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 44.53 | |
| ThanosModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 43.88 | |
| SparseGPTModel=LLaMA-3.2 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 43.78 | |
| LEAPModel=Qwen-2.5 0.5B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 43.41 | |
| MaskLLMModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=2:4 semi-structured2026.05 | 43.39 | |
| SparseGPTModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 43.17 | |
| ADMMModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 43.05 | |
| MaskLLMModel=LLaMA-3.2 1B, Sparsity=50%, Pruning Structure=2:4 semi-structured2026.05 | 42.59 | |
| ThanosModel=LLaMA-3.2 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 42.49 | |
| LEAPModel=Gemma-3 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 42.42 | |
| SparseGPTModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 42.32 | |
| WandaModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 42.29 | |
| ADMMModel=LLaMA-3.2 3B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 42.28 | |
| ThanosModel=LLaMA-3.2 3B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 42.1 | |
| SparseGPTModel=LLaMA-3.2 3B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 41.93 | |
| WandaModel=LLaMA-3.2 3B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 41.71 | |
| ThanosModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 41.62 | |
| LEAPModel=LLaMA-3.2 3B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 41.53 | |
| MaskLLMModel=Qwen-2.5 0.5B, Sparsity=50%, Pruning Structure=2:4 semi-structured2026.05 | 40.91 | |
| ADMMModel=Qwen-2.5 0.5B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 40.41 | |
| WandaModel=Gemma-3 1B, Sparsity=50%, Pruning Structure=Unstructured2026.05 | 39.98 | |
| ThanosModel=Qwen-2.5 0.5B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 38.94 | |
| WandaModel=LLaMA-3.2 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 38.77 | |
| ADMMModel=Gemma-3 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 38.73 | |
| SparseGPTModel=Qwen-2.5 0.5B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 38.65 | |
| ADMMModel=LLaMA-3.2 3B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 38.33 | |
| ThanosModel=LLaMA-3.2 3B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 37.64 | |
| SparseGPTModel=LLaMA-3.2 3B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 37.29 | |
| ThanosModel=Gemma-3 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 36.95 | |
| SparseGPTModel=Gemma-3 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 36.78 | |
| WandaModel=Qwen-2.5 0.5B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 36.64 | |
| WandaModel=LLaMA-3.2 3B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 34.66 | |
| WandaModel=Gemma-3 1B, Sparsity=60%, Pruning Structure=Unstructured2026.05 | 33.83 |