List your company

Stateful Agent Backdoor Attack Evaluation on Note-based instantiation 20 episodes

100Initial Success Rate (R_init)Llama-3.1-8B
94.896.1597.598.85May 7, 2026

Evaluation results

MethodLinks
Llama-3.1-8BModel=Llama-3.1-8B2026.051001009595850105
Qwen2.5-14BModel=Qwen2.5-14B2026.051001009590800205
Ministral-3-14BModel=Ministral-3-14B2026.05100100909080055
Qwen2.5-7BModel=Qwen2.5-7B2026.0595959090900105