List your company

Safety Evaluation on BeaverTails (Harm Score Delta)

0.639Harm Score Delta (Δ)KICK-SETTLE
0.63620.65510.6740.6929May 14, 2026

Evaluation results

MethodLinks
KICK-SETTLEDefense=SDD, Model=Llama-2-7B, Attack Type=Kick-settle2026.050.639
KICK-SETTLEDefense=SDD, Model=Llama-3.1-8B, Attack Type=Kick-settle2026.050.662
KICK-SETTLEDefense=SDD, Model=Qwen3-8B, Attack Type=Kick-settle2026.050.709