Skip to content

A Study on Prompt Injection Attack Against LLM-Integrated Mobile Robotic Systems

Notifications You must be signed in to change notification settings

MoeBuTa/LLMEyesim

Repository files navigation

LLMEyesim

Paper Status

img

Installation

  • Download eyesim
  • create config.yml and set the content according to the config.example.yml
  • set up conda environment
conda env create -f environment.yml
conda activate llmeyesim
pip install -e .
  • Execute the demo
llmeyesim
  • Parameters for V1

    • --world: demo, free, static, dynamic, mixed
    • --model: gpt-4o, gpt-4o-mini
    • --attack: none, omi, ghi
    • --defense: false, true
    • --attack rate: 0.1, 0.3, 0.5, 0.7, 1
  • Parameters for V2

    • TBD

About

A Study on Prompt Injection Attack Against LLM-Integrated Mobile Robotic Systems

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages