Latent Knowledge Elicitation on Deceptive Alignment Benchmark (DAB) 400 scenarios
81.2Elicitation Accuracy
MechELK
Updated 5d ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.04 | 81.2 | |
2026.04 | 79.6 | |
2026.04 | 70.2 | |
2026.04 | 69.8 | |
2026.04 | 68.7 | |
2026.04 | 68.4 | |
2026.04 | 67.3 | |
2026.04 | 67.3 | |
2026.04 | 66.1 | |
2026.04 | 65.9 | |
2026.04 | 62.1 | |
2026.04 | 60.8 |