Safety Alignment on CoSApien Film Production
0.866Safety ScoreLlama-3.1-8B-Inst
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Llama-3.1-8B-InstOptimization Method=Reinforce++2026.05 | 0.866 | 3.493 | 3.113 | |
| CSRMOptimization Method=Reward Distillation2026.05 | 0.83 | 4.625 | 3.83 | |
| CSRMOptimization Method=Reinforce++2026.05 | 0.81 | 4.928 | 3.981 | |
| Llama-3.1-8B-InstOptimization Method=Reward Distillation2026.05 | 0.8 | 4.535 | 3.625 | |
| LlamaGuard-3-8BOptimization Method=Reinforce++2026.05 | 0.778 | 4.275 | 3.391 | |
| LlamaGuard-3-8BOptimization Method=Reward Distillation2026.05 | 0.775 | 4.95 | 3.75 |