Software Engineering Skill Evaluation on SWE-Skills-Bench (slo-implementation)
60Pass Rate (out of 10)S2L
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| S2LSetting=Our method2026.06 | 60 | -0.0251 | 7.96 | |
| Vanilla LLMSetting=Vanilla baseline2026.06 | 40 | — | — | |
| Full Skill TextSetting=Full Skill Text prompting2026.06 | 40 | 0.1566 | 0 |