Software Engineering on SWE tasks Performance Data n = 10 (test)
49.64F1 ScoreIn-Context
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| In-ContextBackbone=Qwen3-8B, Skill Type=In-context skills2026.06 | 49.64 | 87.58 | 33 | |
| No-SkillBackbone=Qwen3-8B, Skill Type=None2026.06 | 48.98 | 87.39 | 37 | |
| ParametricSkillsBackbone=Qwen3-8B, Skill Type=Parametric (Hypernetwork generated)2026.06 | 44.89 | 88.12 | 38.5 | |
| SHINEBackbone=Qwen3-8B, Skill Type=Parametric (SHINE checkpoint)2026.06 | 42.66 | 86.36 | 25 |