Prompt_Inversion-Protection

Research works about prompt inversion attack and prompt copyright protection

Keywords: Prompt Inversion/Stealing/Reconstruction/Extraction

Prompt Inversion Attack

2023

• Prompt Stealing Attacks Against Text-to-Image Generation Models (Xinyue Shen) CISPA Helmholtz Center for Information Security

  • paper: https://arxiv.org/abs/2302.09923 (arXiv)
  • code:

• Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery ()

  • paper: https://arxiv.org/abs/2302.03668 (NIPS 2024)
  • code: https://github.com/YuxinWenRick/hard-prompts-made-easy

• PROPANE: Prompt design as an inverse problem (Rimon Melamed) GWU

  • paper: https://arxiv.org/abs/2311.07064 (arXiv)
  • code: https://github.com/rimon15/propane

• Prompting Hard or Hardly Prompting: Prompt Inversion for Text-to-Image Diffusion Models (Shweta Mahajan) University of British Columbia

  • paper: https://arxiv.org/abs/2312.12416 (arXiv)
  • code:

• Prompts Should not be Seen as Secrets: Systematically Measuring Prompt Extraction Attack Success (Yiming Zhang) Carnegie Mellon University

  • paper: https://arxiv.org/abs/2307.06865 (arXiv)
  • code:

---

Prompt Copyright Protection

2023

• PromptCARE: Prompt Copyright Protection by Watermark Injection and Verification () Zhejiang University

  • paper: https://arxiv.org/abs/2308.02816 (IEEE S&P)
  • code: https://github.com/grasses/PromptCARE

• Secure Your Model: A Simple but Effective Key Prompt Protection Mechanism for Large Language Models ()

  • paper: https://www.researchgate.net/publication/374555007 (ResearchGate)
  • code: