Inference Efficiency and Reasoning on Nine Datasets (Aggregate)
58mAccQwen-3-8B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen-3-8BCompression Ratio=Original2025.10 | 58 | 1 | |
| Llama-3.1-8BCompression Ratio=Original2025.10 | 57 | 1 | |
| NeuronMLPBase Model=Qwen-3-8B, Compression Ratio=0.052025.10 | 56 | 1.18 | |
| NeuronMLPBase Model=Qwen-3-8B, Compression Ratio=0.102025.10 | 55 | 1.28 | |
| NeuronMLPBase Model=Llama-3.1-8B, Compression Ratio=0.052025.10 | 54 | 1.24 | |
| Qwen-3-4BCompression Ratio=Original2025.10 | 54 | 1 | |
| NeuronMLPBase Model=Qwen-3-4B, Compression Ratio=0.052025.10 | 53 | 1.17 | |
| NeuronMLPBase Model=Qwen-3-4B, Compression Ratio=0.102025.10 | 52 | 1.25 | |
| Llama-3.2-3BCompression Ratio=Original2025.10 | 51 | 1 | |
| NeuronMLPBase Model=Qwen-3-8B, Compression Ratio=0.202025.10 | 51 | 1.65 | |
| NeuronMLPBase Model=Llama-3.2-3B, Compression Ratio=0.052025.10 | 49 | 1.22 | |
| NeuronMLPBase Model=Llama-3.1-8B, Compression Ratio=0.102025.10 | 49 | 1.37 | |
| NeuronMLPBase Model=Qwen-3-4B, Compression Ratio=0.202025.10 | 49 | 1.67 | |
| Qwen-3-1.7BCompression Ratio=Original2025.10 | 47 | 1 | |
| NeuronMLPBase Model=Qwen-3-1.7B, Compression Ratio=0.052025.10 | 47 | 1.29 | |
| NeuronMLPBase Model=Qwen-3-1.7B, Compression Ratio=0.102025.10 | 46 | 1.41 | |
| NeuronMLPBase Model=Llama-3.1-8B, Compression Ratio=0.202025.10 | 45 | 1.69 | |
| NeuronMLPBase Model=Llama-3.2-3B, Compression Ratio=0.102025.10 | 44 | 1.34 | |
| NeuronMLPBase Model=Qwen-3-1.7B, Compression Ratio=0.202025.10 | 44 | 1.74 | |
| Llama-3.2-1BCompression Ratio=Original2025.10 | 43 | 1 | |
| NeuronMLPBase Model=Llama-3.2-1B, Compression Ratio=0.052025.10 | 41 | 1.19 | |
| NeuronMLPBase Model=Llama-3.2-3B, Compression Ratio=0.202025.10 | 41 | 1.78 | |
| NeuronMLPBase Model=Llama-3.2-1B, Compression Ratio=0.102025.10 | 38 | 1.28 | |
| NeuronMLPBase Model=Llama-3.2-1B, Compression Ratio=0.202025.10 | 36 | 1.63 |