| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Direct Attack Security | TensorTrust | Direct ASR75.56 | 4 | |
| Overrefusal evaluation | TensorTrust overrefusal | Performance Score91 | 2 | |
| Instruction Hierarchy Robustness | TensorTrust user (dev) | Score91 | 2 | |
| Instruction Hierarchy Robustness | TensorTrust sys-user | Score94 | 2 |