Safety Evaluation on HarmBench Ctx
80.6Refusal RateALIGNBEAM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ALIGNBEAMDraft model=Finance-Llama3-8B, Domain=finance, LBM-DROP=true, alpha=0.5, N=6, K=3, seed=422026.06 | 80.6 | 90.3 | |
| ALIGNBEAMDraft model=Llama-3.1-8B, Domain=general, LBM-DROP=true, alpha=0.5, N=6, K=3, seed=422026.06 | 79 | 80.7 | |
| ALIGNBEAMDraft model=Macro average, LBM-DROP=true, alpha=0.5, K=3, seed=422026.06 | 72.3 | 85.5 | |
| ALIGNBEAMDraft model=Qwen2-7B-Base, Domain=same-vocab ctrl, LBM-DROP=true, alpha=0.5, N=6, K=3, seed=422026.06 | 71 | 90.3 | |
| ALIGNBEAMDraft model=MedLlama-3-8B, Domain=medical, LBM-DROP=true, alpha=0.5, N=6, K=3, seed=422026.06 | 69.4 | 87.1 | |
| ALIGNBEAMDraft model=DeepSeek-Math-7B-Instruct, Domain=math, LBM-DROP=true, alpha=0.5, N=20, K=3, seed=422026.06 | 61.3 | 79 |