List your company

Prompt Injection Attack on Real-world Home (test)

88.3ASRPI3D
64.06870.35976.6582.941Feb 6, 2026

Evaluation results

MethodLinks
PI3DEvaluator MLLM=GPT-4o-mini2026.0288.3
PI3DEvaluator MLLM=Gemini-2.0-Flash2026.0265