List your company

Indirect Prompt Injection on LLM Behavior Subset 1

99.8IRActor-Critic
-3.99222.95449.976.846May 14, 2026

Evaluation results

MethodLinks
Actor-CriticMemory regime=Tool, Model=GPT-5.52026.0599.8
Actor-CriticMemory regime=Tool, Model=GPT-5.42026.0599.4
Actor-CriticMemory regime=Tool, Model=DeepSeek-v42026.0596.2
Actor-CriticMemory regime=Tool, Model=Kimi-K2.62026.0595
Actor-CriticMemory regime=Tool, Model=Gemini-3.12026.0588.6
Actor-CriticMemory regime=External, Model=DeepSeek-v42026.0586.4
Actor-CriticMemory regime=External, Model=GPT-5.42026.0580.4
User ReviewMemory regime=Tool, Model=DeepSeek-v42026.0579
Actor-CriticMemory regime=External, Model=Kimi-K2.62026.0578.8
Actor-CriticMemory regime=External, Model=GPT-5.52026.0575
Actor-CriticMemory regime=Tool, Model=Sonnet-4.62026.0564.2
User ReviewMemory regime=Tool, Model=Kimi-K2.62026.0562.8
User ReviewMemory regime=Tool, Model=Gemini-3.12026.0562.4
Actor-CriticMemory regime=External, Model=Gemini-3.12026.0554.2
Actor-CriticMemory regime=External, Model=Sonnet-4.62026.0513.6
User ReviewMemory regime=External, Model=DeepSeek-v42026.055.4
User ReviewMemory regime=Tool, Model=GPT-5.52026.054.2
User ReviewMemory regime=External, Model=Gemini-3.12026.053.8
User ReviewMemory regime=Tool, Model=GPT-5.42026.053
User ReviewMemory regime=External, Model=GPT-5.52026.052.6
User ReviewMemory regime=External, Model=GPT-5.42026.052.2
User ReviewMemory regime=External, Model=Kimi-K2.62026.052
User ReviewMemory regime=External, Model=Sonnet-4.62026.050.2
User ReviewMemory regime=Tool, Model=Sonnet-4.62026.050