Ignore This Title and HackAPrompt: Exposing Systemic Weaknesses of LLMs through a Global Scale Prompt Hacking Competition

Abstract

Presents results from a global prompt hacking competition with 600K+ adversarial prompts, revealing systemic LLM vulnerabilities across multiple models and defense strategies.

Categories

Framework mappings

@inproceedings{schulhoff2023ignore,
  title = {{Ignore This Title and HackAPrompt: Exposing Systemic Weaknesses of LLMs through a Global Scale Prompt Hacking Competition}},
  author = {Sander Schulhoff and Jeremy Pinto and Anaum Khan and Louis-Francois Bouchard and Chenglei Si and Svetlina Anati and Valen Tagliabue and Anson Liu Kost and Christopher Carnahan and Jordan Boyd-Graber},
  year = {2023},
  month = nov,
  booktitle = {EMNLP 2023},
  eprint = {2311.16119},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2311.16119}
}