Semantic Safety Evaluation on XS-Test CS-RLHF responses
96.43XS ScoreCS-RLHF
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CS-RLHFResponse Source=CS-RLHF2025.10 | 96.43 | 241 | 193 | |
| Human verdictResponse Source=CS-RLHF2025.10 | 91.65 | 225 | 191 | |
| Safe-RLHFResponse Source=CS-RLHF2025.10 | 31.73 | 86 | 51 |