| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| IHEval Conflict Multi-turn | VerIH | Accuracy51.74 | 16 | 25d ago | |
| IHEval Conflict Single-turn | IHDec | Accuracy52.13 | 16 | 25d ago | |
| IHEval Aligned Multi-turn | Accuracy88.8 | 16 | 25d ago | ||
| IHEval Aligned Single-turn | Accuracy84.8 | 16 | 25d ago | ||
| Instruction Hierarchy Evaluation Conflicts Between Message Types | Developer <> User Conflict Rate68 | 2 | 2mo ago |