Sabotage detection on BigCodeBench Sabotage (reasoning LLM attacker)
0.87log-AUROC
Extract-and-Evaluate
Updated 4d ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.01 | 0.87 | - | |
2026.01 | 0.73 | - | |
2026.01 | 0.73 | - | |
2026.01 | 0.61 | - | |
2026.01 | 0.58 | - | |
2026.01 | 0.55 | - | |
2026.01 | 0.48 | - | |
2026.01 | 0.45 | - |