Research works about prompt inversion attack and prompt copyright protection
Keywords: Prompt Inversion/Stealing/Reconstruction/Extraction
-
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:
-
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: