Question Answering on GPQA Diamond (GPQA metric)
69.19GPQA ScoreOriginal
Evaluation Results
| Method | Links | |
|---|---|---|
| OriginalModel=Seed-OSS-36B-Instruct, MLP Sparsity=0%2026.06 | 69.19 | |
| SpenseGPT+Model=Seed-OSS-36B-Instruct, MLP Sparsity=25%2026.06 | 68.18 | |
| SpenseGPTModel=Seed-OSS-36B-Instruct, MLP Sparsity=25%2026.06 | 66.84 | |
| SpenseGPT+Model=Qwen3-32B, MLP Sparsity=25%2026.06 | 66.83 | |
| SpenseGPT+Model=Seed-OSS-36B-Instruct, MLP Sparsity=37.5%2026.06 | 66.5 | |
| OriginalModel=Qwen3-32B, MLP Sparsity=0%2026.06 | 65.49 | |
| SpenseGPTModel=Seed-OSS-36B-Instruct, MLP Sparsity=37.5%2026.06 | 64.31 | |
| SpenseGPTModel=Qwen3-32B, MLP Sparsity=25%2026.06 | 62.46 | |
| SpenseGPT+Model=Qwen3-32B, MLP Sparsity=37.5%2026.06 | 61.45 | |
| SparseGPTModel=Seed-OSS-36B-Instruct, MLP Sparsity=50%2026.06 | 60.27 | |
| SpenseGPTModel=Qwen3-32B, MLP Sparsity=37.5%2026.06 | 59.43 | |
| SparseGPTModel=Qwen3-32B, MLP Sparsity=50%2026.06 | 53.71 | |
| HyperPruneModel=Qwen3-32B, MLP Sparsity=50%2026.06 | 23.57 | |
| WandaModel=Qwen3-32B, MLP Sparsity=50%2026.06 | 15.49 | |
| HyperPruneModel=Seed-OSS-36B-Instruct, MLP Sparsity=50%2026.06 | 7.58 | |
| WandaModel=Seed-OSS-36B-Instruct, MLP Sparsity=50%2026.06 | 6.9 |