Skip to content
Search
paperJanuary 2024ReviewedOpen access

Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Models

Jingwei Yi, Yueqi Xie, Bin Zhu, Keegan Hines, Emre Kiciman, Guangzhong Sun, Xing Xie, Fangzhao Wu

arXiv preprint

View resourcePDF

Abstract

Provides a benchmark for indirect prompt injection attacks and evaluates several defense strategies including perplexity-based detection and sandwich defense.

Categories

Prompt InjectionInput FilteringBenchmarks & Evaluation
#indirect-injection#benchmark#defense-evaluation

Framework mappings

OWASP Top 10 for LLM Applications
  • LLM01Prompt Injection
MITRE ATLAS
  • AML.T0051LLM Prompt Injection

Cite

@misc{yi2024benchmarking,
  title = {{Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Models}},
  author = {Jingwei Yi and Yueqi Xie and Bin Zhu and Keegan Hines and Emre Kiciman and Guangzhong Sun and Xing Xie and Fangzhao Wu},
  year = {2024},
  month = jan,
  eprint = {2312.14197},
  archivePrefix = {arXiv},
  doi = {10.1145/3690624.3709179},
  url = {https://arxiv.org/abs/2312.14197}
}