List your company

LLM Safety Defense on CoSafe

15.42ASRMTSA-T3
14.388821.349428.3135.2706May 22, 2025

Evaluation results

MethodLinks
MTSA-T3Target Model=Llama2-7B-Chat, Iteration=32025.0515.42
MTSA-T3Target Model=Zephyr-7B-Beta, Iteration=32025.0518.78
BaselineTarget Model=Llama2-7B-Chat2025.0539.4
BaselineTarget Model=Zephyr-7B-Beta2025.0541.2