Compositional Jailbreaking: An Empirical Analysis of Mutator Chain Interactions in Aligned LLMs

Abstract

Jailbreaking attacks on large language models pose a significant threat to AI safety by enabling the generation of harmful or restricted content. While prior work has explored both handcrafted and automated jailbreak strategies, the potential for compositional interaction between simple attacks remains underexplored. This paper presents a systematic study of mutator chaining, in which weak jailbreak transformations are applied sequentially to characterize how they interact: whether they reinforc

Categories

Framework mappings

Suggested from the entry's categories.

@misc{bugnot2026compositional,
  title = {{Compositional Jailbreaking: An Empirical Analysis of Mutator Chain Interactions in Aligned LLMs}},
  author = {Reinelle Jan Bugnot and Soohyeon Choi and Hoon Wei Lim and Yue Duan},
  year = {2026},
  month = may,
  eprint = {2605.15598},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.15598}
}