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

Prarabdh Shukla, Ritik, Suhas Rao, Arpit Agarwal, Arjun Bhagoji · 2026

Read Jailbreaking for the Average Jane : Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries by Prarabdh Shukla, Ritik, Suhas Rao, Arpit Agarwal, Arjun Bhagoji (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-26936.txt.

Introduction

Prarabdh Shukla, Ritik, Suhas Rao, Arpit Agarwal, and Arjun Bhagoji test whether an ordinary non-expert could jailbreak a language model with public tools. Jailbreaks, prompts that bypass safety rules, are now widely known. A would-be attacker needs a strong jailbreak for their target and a well-crafted harmful query. The authors supply both halves systematically.

A bandit algorithm learns which jailbreak works best. It explores options on a few queries, then exploits the winner at scale, like a gambler favouring the best slot machine. Their FrankensteinBench merges over eleven thousand harmful queries from seven sources, labelled by the skill needed to craft them. The concern proved justified: bandit-chosen attacks reached very high average success rates across fifteen leading open models. The authors present the work as a warning about accessible misuse.

Catalog record

CREATORPrarabdh Shukla, Ritik, Suhas Rao, Arpit Agarwal, Arjun Bhagoji
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE136 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.26936

Browse the full library · About the Tower of Babel