Jailbreaking for the Average Jane: Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries

Abstract

With a profusion of jailbreaks for LLMs now widely known, a growing concern is that non-expert malicious actors ("the average Jane") could elicit actionable responses to malicious requests. In this work, we examine whether this concern is justified. A non-expert malicious actor requires two ingredients for a successful attack: a powerful jailbreak for their target model, acting on an effective malicious query. For the former, we propose a novel attack strategy based on the multi-armed bandit fra

Categories

Framework mappings

Suggested from the entry's categories.

@misc{shukla2026jailbreaking,
  title = {{Jailbreaking for the Average Jane: Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries}},
  author = {Prarabdh Shukla and Ritik and Suhas Rao and Arpit Agarwal and Arjun Bhagoji},
  year = {2026},
  month = jun,
  eprint = {2606.26936},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2606.26936}
}