Disentangling Adversarial Prompts: A Semantic-Graph Defense for Robust LLM Security

Abstract

Large Language Models (LLMs) are increasingly vulnerable to adversarial prompts that exploit semantic ambiguities to bypass safety mechanisms, resulting in harmful or inappropriate outputs. Such attacks, including jailbreaking and prompt injection, pose significant risks to the integrity and availability of LLMs in security-critical applications. This paper proposes the Adversarial Prompt Disentanglement (APD) framework, a novel defense mechanism that proactively identifies and neutralizes malic

Categories

Framework mappings

Suggested from the entry's categories.

@misc{fang2026disentangling,
  title = {{Disentangling Adversarial Prompts: A Semantic-Graph Defense for Robust LLM Security}},
  author = {Xiang Fang and Wanlong Fang},
  year = {2026},
  month = may,
  eprint = {2605.27823},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.27823}
}