Sycophancy Mitigation on TruthfulQA
25Sycophancy Rate
Linear Probe MHA
Updated 4d ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.01 | 25 | |
2026.01 | 34.4 | |
2026.01 | 37.5 | |
2026.01 | 40.7 | |
2026.01 | 40.7 | |
2026.01 | 41.2 | |
2026.01 | 42.6 | |
2026.01 | 42.7 | |
2026.01 | 43.9 | |
2026.01 | 44.2 | |
2026.01 | 44.4 | |
2026.01 | 44.7 | |
2026.01 | 45.1 | |
2026.01 | 51.7 |