Double-Gaming: Jailbreak Attacks against LLM based on Red-Blue Team Game Theory

Abstract

Despite existing security alignment mechanisms, Large Language Models (LLMs) remain vulnerable to jailbreak attacks under static defenses. To address this, we propose a novel jailbreak attack and defense optimization framework based on Red-Blue Team dynamic game theory. This framework establishes an automated adversarial mechanism between red and blue teams, achieving closed-loop optimization of instruction generation, semantic interception, and strategy evolution. The red team employs reinforce

Categories

Framework mappings

Suggested from the entry's categories.

@inproceedings{ma2026doublegaming,
  title = {{Double-Gaming: Jailbreak Attacks against LLM based on Red-Blue Team Game Theory}},
  author = {Chenlu Ma and Huairui Zhao and G. Nie and Baiyang Ji and Beibei Li and Haibin Zheng},
  year = {2026},
  month = apr,
  booktitle = {2026 8th International Conference on Software Engineering and Computer Science (CSECS)},
  doi = {10.1109/CSECS69124.2026.11541892},
  url = {https://www.semanticscholar.org/paper/17d55e7e6983830e40b9cb742d612e16687b8365}
}