Safety and Cultural Bias Evaluation on KoBBQ
89.18AccuracyKanana-1.5-8B-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| Kanana-1.5-8B-InstructStage=Base2026.06 | 89.18 | |
| Kanana-1.5-8B-InstructStage=Post-trained2026.06 | 88.9 | |
| EXAONE-3.5-7.8B-InstructStage=Post-trained2026.06 | 85.87 | |
| EXAONE-3.5-7.8B-InstructStage=Base2026.06 | 85.78 | |
| Qwen-2.5-7B-InstructStage=Post-trained2026.06 | 83.29 | |
| Qwen-2.5-7B-InstructStage=Base2026.06 | 82.99 | |
| A.X-4.0-LightStage=Post-trained2026.06 | 77.39 | |
| Llama-3.1-8B-InstructStage=Post-trained2026.06 | 75.74 | |
| A.X-4.0-LightStage=Base2026.06 | 72.14 | |
| Llama-3.1-8B-InstructStage=Base2026.06 | 71.33 | |
| Gemma-3-4B-ITStage=Post-trained2026.06 | 53.36 | |
| Gemma-3-4B-ITStage=Base2026.06 | 53.29 |