List your company

Software Engineering on SWE tasks Robustness Fix n = 55 (test)

57.9F1 ScoreParametricSkills
40.594445.087249.5854.0728Jun 29, 2026

Evaluation results

MethodLinks
ParametricSkillsBackbone=Qwen3-8B, Skill Type=Parametric (Hypernetwork generated)2026.0657.990.1775.09
In-ContextBackbone=Qwen3-8B, Skill Type=In-context skills2026.0651.1388.8366
No-SkillBackbone=Qwen3-8B, Skill Type=None2026.0650.7989.3766.91
SHINEBackbone=Qwen3-8B, Skill Type=Parametric (SHINE checkpoint)2026.0641.2688.5661