Safety Evaluation on Norwegian pack (full 36-scenario)
43.68Judge-L ScoreBorealis-27B
Evaluation Results
| Method | Links | |
|---|---|---|
| Borealis-27BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 43.68 | |
| Borealis-12BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 42.29 | |
| Gemma-3-27BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 37.64 | |
| Gemma-3-12BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 35.28 | |
| Borealis-4BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 27.91 | |
| Gemma-3-4BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 13.74 | |
| Borealis-1BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 4.58 | |
| Gemma-3-1BJudge=Qwen3.5-122B-A10B (Q4_K_S), n (sample size)=10, Number of scenarios=362026.05 | 3.82 |