List your company

Adversarial Robustness on Harmbench 39 standard behavior examples

0Attack Success RateZEPHYR-CAT
-4235077May 24, 2024

Evaluation results

MethodLinks
ZEPHYR-CATBackbone=Zephyr, Attack=Simple Adaptive2024.050
PHI-3-MINI-2B-CATBackbone=Phi-3-Mini-2B, Attack=Simple Adaptive2024.050
PHI-3-MINI-2B-CAPOBackbone=Phi-3-Mini-2B, Attack=Simple Adaptive2024.050
PHI-3-MINIBackbone=Phi-3-Mini, Attack=Simple Adaptive2024.05100