False Alarm Detection on IFEval
0False Alarm Rateembedding disruption method
Evaluation Results
| Method | Links | |
|---|---|---|
| embedding disruption methodTarget Model=Vicuna2026.05 | 0 | |
| embedding disruption methodTarget Model=Qwen2.52026.05 | 0 | |
| embedding disruption methodTarget Model=Llama22026.05 | 2 | |
| Llama-Guard-3Target Model=Vicuna2026.05 | 4 | |
| Llama-Guard-3Target Model=Llama22026.05 | 4 | |
| Llama-Guard-3Target Model=Qwen2.52026.05 | 4 | |
| GuardReasonerTarget Model=Vicuna2026.05 | 24 | |
| GuardReasonerTarget Model=Llama22026.05 | 24 | |
| GuardReasonerTarget Model=Qwen2.52026.05 | 24 |