Misalignment Detection on Sycophancy Bloom behaviors (in-distribution)
93.5AUROC
RepE baseline
Updated 1mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.06 | 93.5 | |
2026.06 | 93 | |
2026.06 | 92.8 | |
2026.06 | 91.3 |
RepE baseline
| Method | Links | |
|---|---|---|
2026.06 | 93.5 | |
2026.06 | 93 | |
2026.06 | 92.8 | |
2026.06 | 91.3 |