Zero-shot Learning on 7 Downstream Tasks Avg
74.1Average ScoreFP16
Evaluation results
| Method | Links | |
|---|---|---|
| FP16Model Architecture=Qwen 3-14B, Rank=-2026.03 | 74.1 | |
| GlowQModel Architecture=Qwen 3-14B2026.03 | 73.84 | |
| ZeroQuant-V2Model Architecture=LLaMA 3.1-8B2026.03 | 73.48 | |
| GlowQModel Architecture=LLaMA 3.1-8B2026.03 | 73.33 | |
| FP16Model Architecture=LLaMA 3.1-8B, Rank=-2026.03 | 73.29 | |
| L2QERModel Architecture=Qwen 3-14B, Rank=642026.03 | 73.24 | |
| GlowQ-SModel Architecture=Qwen 3-14B2026.03 | 73.24 | |
| QERAModel Architecture=Qwen 3-14B2026.03 | 73.14 | |
| QERAModel Architecture=LLaMA 3.1-8B2026.03 | 72.86 | |
| ZeroQuant-V2Model Architecture=Qwen 3-14B2026.03 | 72.62 | |
| GlowQ-SModel Architecture=LLaMA 3.1-8B2026.03 | 72.62 | |
| L2QERModel Architecture=LLaMA 3.1-8B, Rank=642026.03 | 72.43 | |
| FP16Model Architecture=Qwen 3-8B, Rank=-2026.03 | 71.48 | |
| GlowQModel Architecture=Qwen 3-8B2026.03 | 70.71 | |
| GlowQ-SModel Architecture=Qwen 3-8B2026.03 | 70.29 | |
| ZeroQuant-V2Model Architecture=Qwen 3-8B2026.03 | 70.19 | |
| QERAModel Architecture=Qwen 3-8B2026.03 | 69.86 | |
| L2QERModel Architecture=Qwen 3-8B, Rank=642026.03 | 69.52 | |
| FP16Model Architecture=LLaMA 3.2-3B, Rank=-2026.03 | 67.14 | |
| GlowQModel Architecture=LLaMA 3.2-3B2026.03 | 66.9 | |
| GlowQ-SModel Architecture=LLaMA 3.2-3B2026.03 | 66.33 | |
| L2QERModel Architecture=LLaMA 3.2-3B, Rank=642026.03 | 66.19 | |
| QERAModel Architecture=LLaMA 3.2-3B2026.03 | 65.48 | |
| ZeroQuant-V2Model Architecture=LLaMA 3.2-3B2026.03 | 65.38 | |
| O-FANGModel=LLaMA-3.1-8B, Sparsity=20%2025.12 | 64.53 | |
| OBCModel=LLaMA-3.1-8B, Sparsity=20%2025.12 | 56.88 | |
| O-FANGModel=Qwen2.5-7B, Sparsity=40%2025.12 | 52.14 | |
| OBCModel=Qwen2.5-7B, Sparsity=40%2025.12 | 48.82 |