Question Answering on CM17k
39.75AccuracyDGAO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DGAOModel=Qwen-2.5-14B2026.05 | 39.75 | 81.73 | 47.36 | |
| SFTModel=Qwen-2.5-14B2026.05 | 38.87 | 79.97 | 51.27 | |
| PAFTModel=Qwen-2.5-14B2026.05 | 29.76 | 85.08 | 64.58 | |
| DGAOModel=Qwen-2.5-3B2026.05 | 29.35 | 77.24 | 46.99 | |
| SFTModel=Qwen-2.5-3B2026.05 | 28.33 | 78.02 | 51.24 | |
| DGAOModel=Llama-3.2-11B2026.05 | 21.87 | 72.56 | 51.58 | |
| SFTModel=Llama-3.2-11B2026.05 | 21.8 | 72.38 | 56.85 | |
| PAFTModel=Qwen-2.5-3B2026.05 | 18.83 | 83.2 | 64.94 | |
| PAFTModel=Llama-3.2-11B2026.05 | 17.03 | 75.92 | 65.62 | |
| DGAOModel=Llama-3.2-3B2026.05 | 15.19 | 70.61 | 51.14 | |
| DGAOBackbone=Llama-3.2-3B2026.05 | 15.19 | 70.61 | 51.14 | |
| DEmOBackbone=Llama-3.2-3B2026.05 | 15.15 | 70.51 | 56.71 | |
| SFTModel=Llama-3.2-3B2026.05 | 15.08 | 70.45 | 56.8 | |
| SFTBackbone=Llama-3.2-3B2026.05 | 15.08 | 70.45 | 56.8 | |
| SPHSBackbone=Llama-3.2-3B2026.05 | 14.99 | 70.36 | 57.17 | |
| PromptBackbone=Llama-3.2-3B2026.05 | 14.93 | 70.68 | 57.03 | |
| GlobalEBackbone=Llama-3.2-3B2026.05 | 14.88 | 70.49 | 57.06 | |
| DGAOModel=Gemma-3-4B2026.05 | 11.78 | 59.57 | 47.12 | |
| SFTModel=Gemma-3-4B2026.05 | 11.39 | 59.32 | 49.26 | |
| PAFTModel=Llama-3.2-3B2026.05 | 10 | 77.12 | 67.96 | |
| PAFTModel=Gemma-3-4B2026.05 | 9.54 | 70.18 | 52.84 | |
| PINEBackbone=Llama-3.2-3B2026.05 | 7.92 | 62.25 | 56.88 |