paper-with-me

홈 › Papers

Jailbreak Distillation: Renewable Safety Benchmarking

2025-05-28 · Jingyu Zhang, Ahmed Elgohary, Xiawei Wang, A S M Iftekhar, Ahmed Magooda, Benjamin Van Durme, Daniel Khashabi, Kyle Jackson

Large language models (LLMs) are rapidly deployed in critical applications, raising urgent needs for robust safety benchmarking. We propose Jailbreak Distillation (JBDistill), a novel benchmark construction framework that "distills" jailbreak attacks into high-quality and easily-updatable safety benchmarks. JBDistill utilizes a small set of development models and existing jailbreak attack algorithms to create a candidate prompt pool, then employs prompt selection algorithms to identify an effective subset of prompts as safety benchmarks. JBDistill addresses challenges in existing safety evaluation: the use of consistent evaluation prompts across models ensures fair comparisons and reproducibility. It requires minimal human effort to rerun the JBDistill pipeline and produce updated benchmarks, alleviating concerns on saturation and contamination. Extensive experiments demonstrate our benchmarks generalize robustly to 13 diverse evaluation models held out from benchmark construction, including proprietary, specialized, and newer-generation LLMs, significantly outperforming existing safety benchmarks in effectiveness while maintaining high separability and diversity. Our framework thus provides an effective, sustainable, and adaptable solution for streamlining safety evaluation.

📄 PDF Abstract BibTeX arXiv:2505.22037

Code (0)

등록된 구현이 없습니다.

Tasks

BenchmarkingDiversity

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety

2025-12-08 · Max Zhang, Derek Liu, Kai Zhang, Joshua Franco 외 arxiv

Large language models (LLMs) are increasingly deployed worldwide, yet their safety alignment remains predominantly English-centric. This allows for vulnerabilities in non-English contexts, especially with low-resource la…

parameter-efficient fine-tuningKnowledge Distillation

h4rm3l: A language for Composable Jailbreak Attack Synthesis

2024-08-09 · Moussa Koulako Bala Doumbouya, Ananjan Nandi, Gabriel Poesia, Davide Ghilardi 외

Despite their demonstrated valuable capabilities, state-of-the-art (SOTA) widely deployed large language models (LLMs) still have the potential to cause harm to society due to the ineffectiveness of their safety filters,…

BenchmarkingProgram SynthesisRed Teaming

Multilingual Safety Alignment via Self-Distillation

2026-05-03 · Ruiyang Qin, Qingzhuo Wang, Dongrui Liu, Qiang Li 외 arxiv

Large language models (LLMs) exhibit severe multilingual safety misalignment: they possess strong safeguards in high-resource languages but remain highly vulnerable to jailbreak attacks in low-resource languages. Current…

The Struggle Between Continuation and Refusal: A Mechanistic Analysis of the Continuation-Triggered Jailbreak in LLMs

2026-03-09 · Yonghong Deng, Zhen Yang, Ping Jian, Xinyue Zhang 외 arxiv

With the rapid advancement of large language models (LLMs), the safety of LLMs has become a critical concern. Despite significant efforts in safety alignment, current LLMs remain vulnerable to jailbreaking attacks. Howev…

Behind the Mask: Benchmarking Camouflaged Jailbreaks in Large Language Models

2025-09-05 · Youjia Zheng, Mohammad Zandsalimy, Shanu Sushmita arxiv

Large Language Models (LLMs) are increasingly vulnerable to a sophisticated form of adversarial prompting known as camouflaged jailbreaking. This method embeds malicious intent within seemingly benign language to evade e…