Semantic Intent Recognition on XSTest
96.43XS Test ScoreCS-RLHF
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CS-RLHF2025.10 | 96.43 | 241 | 193 | |
| Human Verdict2025.10 | 91.65 | 225 | 191 | |
| CS-RLHFEvaluation setting=Using Safe-RLHF-Generated Responses2025.10 | 91.05 | 225 | 187 | |
| Human VerdictEvaluation setting=Using Safe-RLHF-Generated Responses2025.10 | 88.73 | 221 | 179 | |
| Safe-RLHF2025.10 | 31.73 | 86 | 51 | |
| Safe-RLHFEvaluation setting=Using Safe-RLHF-Generated Responses2025.10 | 7.14 | 18 | 14 |