Multilingual Safety on MultiJail In-Distribution Languages (test)
33.65Safety Score (EN)SDRRL
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SDRRLBackbone=Gemma-2-9B-it2026.02 | 33.65 | 16.19 | 18.41 | 17.78 | 14.92 | 20.19 | |
| RawBackbone=Qwen 2.5-7B-Instruct2026.02 | 7.3 | 4.76 | 10.48 | 33.02 | 30.79 | 17.27 | |
| SDRRLBackbone=Qwen 2.5-7B-Instruct2026.02 | 6.98 | 5.39 | 9.84 | 30.79 | 31.42 | 16.88 | |
| MPOBackbone=Qwen 2.5-7B-Instruct2026.02 | 6.35 | 3.81 | 4.76 | 5.71 | 0.32 | 4.19 | |
| DPOBackbone=Qwen 2.5-7B-Instruct2026.02 | 2.86 | 1.9 | 5.08 | 17.78 | 25.08 | 10.54 | |
| MPOBackbone=Gemma-2-9B-it2026.02 | 2.22 | 4.44 | 2.22 | 6.98 | 2.54 | 3.68 | |
| RawBackbone=Gemma-2-9B-it2026.02 | 1.27 | 4.13 | 2.54 | 6.98 | 6.98 | 4.38 | |
| DPO+MLCBackbone=Qwen 2.5-7B-Instruct2026.02 | 0.63 | 0.32 | 0.95 | 0.95 | 0.63 | 0.7 | |
| DPO+MLCBackbone=Gemma-2-9B-it2026.02 | 0.63 | 0 | 0.63 | 0.32 | 0 | 0.32 | |
| DPOBackbone=Gemma-2-9B-it2026.02 | 0 | 2.22 | 1.9 | 5.4 | 4.13 | 2.73 |