Physical Commonsense Reasoning on PIQA (Accuracy (%), Avg.)
81.08Accuracy (%)MoNE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MoNEType=P25% Q4b, Storage=6.00GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 81.08 | 63.97 | |
| He et al.Type=P25% Q4b, Storage=6.00GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 81.08 | 63.97 | |
| BaselineType=None, Storage=32.7GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 80.52 | 65.62 | |
| BaselineType=None, Storage=31.41GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 80.2 | 68.6 | |
| MoNEType=P25%, Storage=24.00GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 79.76 | 65.65 | |
| WandaType=P25%, Storage=31.41GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 79.65 | 66.62 | |
| OursQType=P25% Q4b, Storage=6.26GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 79.6 | 63.78 | |
| EAC-MoEType=P11% Q3.03b, Storage=7.19GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 79.54 | 65.3 | |
| OursQType=P25% Q4b, Storage=6.00GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 79.54 | 67.61 | |
| WandaType=P25% Q4b, Storage=7.70GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 79.38 | 69.2 | |
| PuzzleMoEType=M25%, Storage=25.04GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 78.7 | 62.36 | |
| WandaType=P25%, Storage=32.7GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 78.51 | 61.56 | |
| He et al.Type=P25% Q4b, Storage=7.70GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 78.5 | 59.7 | |
| PuzzleMoEType=M50%, Storage=17.34GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 78.4 | 62.08 | |
| Attribution-guided and Coverage-Maximized Expert-wise PruningType=P50%, Storage=16.57GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 78.4 | 64.92 | |
| He et al.Type=P50%, Storage=16.57GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 78.2 | 53.88 | |
| Attribution-guided and Coverage-Maximized Expert-wise PruningType=P50%, Storage=17.34GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 77.97 | 62.34 | |
| WandaType=P50%, Storage=32.7GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 77.86 | 60.82 | |
| C-PruneType=P25% Q4b, Storage=6.00GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 77.86 | 61.81 | |
| EAC-MoEType=P38% Q3.03b, Storage=4.47GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 77.2 | 62.7 | |
| FPModel=Dream-Instruct, Quantization=Full Precision2026.06 | 75.79 | 71.01 | |
| FPModel=Dream-Base, Quantization=Full Precision2026.06 | 75.41 | 70.01 | |
| MoNEType=P50%, Storage=16.57GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 75.3 | 59.8 | |
| FAIR-CalibModel=Dream-Instruct, Quantization=W4A42026.06 | 73.12 | 66.66 | |
| FAIR-CalibModel=Dream-Base, Quantization=W4A42026.06 | 73.07 | 64.64 | |
| QuaRotModel=Dream-Base, Quantization=W4A42026.06 | 72.52 | 57.49 | |
| FlatQuantModel=Dream-Instruct, Quantization=W4A42026.06 | 71.82 | 63.98 | |
| FlatQuantModel=Dream-Base, Quantization=W4A42026.06 | 71.65 | 62.08 | |
| QuaRotModel=Dream-Instruct, Quantization=W4A42026.06 | 71.11 | 61.14 | |
| He et al.Type=P50%, Storage=17.34GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 68.17 | 44.17 | |
| WandaType=P50%, Storage=31.41GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 64.15 | 49.29 | |
| RTNModel=Dream-Instruct, Quantization=W4A42026.06 | 63.33 | 45.76 | |
| RTNModel=Dream-Base, Quantization=W4A42026.06 | 61.7 | 42 | |
| C-PruneType=P50%, Storage=16.57GB, Model Family=DeepSeek-V2-Lite, Evaluation Protocol=Zero-shot2026.06 | 58.92 | 42.39 | |
| MoNEType=P25%, Storage=25.04GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 53.37 | 33.94 | |
| MoNEType=P50%, Storage=17.34GB, Model Family=Deepseek-MoE-16B, Evaluation Protocol=Zero-shot2026.06 | 49.51 | 33.24 |