List your company
ResearchFormerly WizWand

Data Extraction Attack on EHRAgent

83Equality (EQ)ADAM
3.9624.484565.52Apr 10, 2026
Updated 6mo ago

Evaluation results

MethodLinks
ADAMModel=ChatGPT-4, Number of prompts=30, Memory size=3002026.04839297100
ADAMModel=Qwen2-72B, Number of prompts=30, Memory size=3002026.04819093100
ADAMModel=Mistral-7B-Instruct, Number of prompts=30, Memory size=3002026.04808983100
ADAMModel=Llama2-7b-chat, Number of prompts=30, Memory size=3002026.04778593100
PirateModel=ChatGPT-4, Number of prompts=30, Memory size=3002026.04596630100
PirateModel=Qwen2-72B, Number of prompts=30, Memory size=3002026.04556123100
MEXTRAModel=ChatGPT-4, Number of prompts=30, Memory size=3002026.0455615090
MEXTRAModel=Qwen2-72B, Number of prompts=30, Memory size=3002026.0452584488
MEXTRAModel=Mistral-7B-Instruct, Number of prompts=30, Memory size=3002026.0446513690
MEXTRAModel=Llama2-7b-chat, Number of prompts=30, Memory size=3002026.0444493889
PirateModel=Llama2-7b-chat, Number of prompts=30, Memory size=3002026.0440441593
RAG-ThiefModel=Qwen2-72B, Number of prompts=30, Memory size=3002026.04394310100
PirateModel=Mistral-7B-Instruct, Number of prompts=30, Memory size=3002026.04353920100
RAG-ThiefModel=ChatGPT-4, Number of prompts=30, Memory size=3002026.0432360100
RAG-ThiefModel=Llama2-7b-chat, Number of prompts=30, Memory size=3002026.0431340100
RAG-ThiefModel=Mistral-7B-Instruct, Number of prompts=30, Memory size=3002026.042528083
VanillaModel=ChatGPT-4, Number of prompts=30, Memory size=3002026.041415047
VanillaModel=Qwen2-72B, Number of prompts=30, Memory size=3002026.041011033
VanillaModel=Llama2-7b-chat, Number of prompts=30, Memory size=3002026.0489026
VanillaModel=Mistral-7B-Instruct, Number of prompts=30, Memory size=3002026.0478023