Paired Vulnerability Detection on PRIMEVUL paired (val)
22.7P-CRandom Baseline
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Random Baseline2025.07 | 22.7 | 26.24 | 26.42 | 24.65 | |
| GPT-4Method=Chain-of-Thought2025.07 | 12.94 | 54.26 | 24.47 | 8.33 | |
| White-BasiliskMethod=Fine-tuning2025.07 | 12.92 | 42.08 | 42.92 | 2.08 | |
| GPT-4Method=Two-shot2025.07 | 5.14 | 71.63 | 21.45 | 1.77 | |
| CodeGen2.5Method=Fine-tuning2025.07 | 3.01 | 10.82 | 84.22 | 1.95 | |
| StarCoder2-7BMethod=Fine-tuning2025.07 | 2.3 | 8.16 | 88.3 | 1.24 | |
| CodeBERTMethod=Fine-tuning2025.07 | 1.77 | 11.35 | 86.17 | 0.71 |