| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Social Sycophancy (Held-out) | Final SYPR | AUROC91.9 | 16 | 1mo ago | |
| Anthropic Sycophancy Dataset | Contrastive + LR | Accuracy98.3 | 14 | 1mo ago | |
| Culture Dataset | CLiF + SVM | Accuracy98.3 | 12 | 1mo ago | |
| Office Scenarios Dataset | CLiF + SVM | Accuracy98.3 | 12 | 1mo ago | |
| Non-US Policy Dataset | CLiF + SVM | Accuracy100 | 12 | 1mo ago | |
| Sycophancy benchmark (full evaluation set) | Hypocrisy Gap | AUROC0.732 | 12 | 5mo ago | |
| LURE Truthfulness | Sycophancy Rate10 | 10 | 2mo ago | ||
| SYCON-Bench raw n=100 | Sycophancy Rate68 | 10 | 2mo ago |