Aggregate Performance Evaluation on Average SST2, AGNEWS, GSM8K
26.47HS ScoreSFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SFTBackbone=Qwen-2-7B2026.02 | 26.47 | 81.25 | |
| VaccineBackbone=Qwen-2-7B2026.02 | 25.54 | 80.12 | |
| LisaBackbone=Qwen-2-7B2026.02 | 12.54 | 80.61 | |
| AntibodyModel=Gemma-2-9B2026.02 | 4.71 | 79.16 | |
| BoosterBackbone=Qwen-2-7B2026.02 | 3.21 | 82.65 | |
| Booster + LisaModel=Gemma-2-9B2026.02 | 2.72 | 78.71 | |
| Booster + LisaModel=Llama-2-7B2026.02 | 1.84 | 62.57 | |
| Booster + LisaModel=Qwen-2-7B2026.02 | 1.67 | 82.98 | |
| AntibodyModel=Llama-2-7B2026.02 | 1.32 | 65.24 | |
| AntibodyBackbone=Qwen-2-7B2026.02 | 0.72 | 82.95 | |
| AntibodyModel=Qwen-2-7B2026.02 | 0.72 | 82.95 |