Activation-Guided Local Editing for Jailbreaking Attacks

TL;DR

This paper introduces AGILE, a two-stage activation-guided local editing framework for jailbreaking language models, achieving high success rates and transferability while resisting defenses.

Contribution

It presents a novel two-stage method combining scenario-based rephrasing and hidden state-guided edits to improve jailbreak effectiveness.

Findings

Achieves up to 37.74% higher attack success rate over baselines.

Demonstrates strong transferability to black-box models.

Remains effective against prominent defense mechanisms.

Abstract

Jailbreaking is an essential adversarial technique for red-teaming these models to uncover and patch security flaws. However, existing jailbreak methods face significant drawbacks. Token-level jailbreak attacks often produce incoherent or unreadable inputs and exhibit poor transferability, while prompt-level attacks lack scalability and rely heavily on manual effort and human ingenuity. We propose a concise and effective two-stage framework that combines the advantages of these approaches. The first stage performs a scenario-based generation of context and rephrases the original malicious query to obscure its harmful intent. The second stage then utilizes information from the model's hidden states to guide fine-grained edits, effectively steering the model's internal representation of the input from a malicious toward a benign one. Extensive experiments demonstrate that this method achieves state-of-the-art Attack Success Rate, with gains of up to 37.74% over the strongest baseline, and exhibits excellent transferability to black-box models. Our analysis further demonstrates that AGILE maintains substantial effectiveness against prominent defense mechanisms, highlighting the limitations of current safeguards and providing valuable insights for future defense development. Our code is available at https://github.com/SELGroup/AGILE.