Refusal Detection on XSTest Refusal OOD (test)
100MC RecallACS Refusal Probe
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ACS Refusal ProbeTarget Model=Phi2026.05 | 100 | 96 | |
| ACS Refusal ProbeTarget Model=Llama2026.05 | 99 | 97 | |
| ACS Refusal ProbeTarget Model=Qwen2026.05 | 99 | 96 | |
| ACS Refusal ProbeTarget Model=LQMP (4-fam)2026.05 | 99 | 96 | |
| ACS Refusal ProbeTarget Model=LQMPG (5-fam)2026.05 | 99 | 93 | |
| ACS Refusal ProbeTarget Model=Mistral2026.05 | 98 | 97 | |
| ACS Refusal ProbeTarget Model=Gemma2026.05 | 97 | 78 |