Fetching the paper…

Reading the bibliography…

2023

Jailbreaker in Jail: Moving Target Defense for Large Language Models

Chen, Bocheng, Paliwal, Advait, Yan, Qiben

Original
PDF
Post on X

Understand

Large language models (LLMs), known for their capability in understanding and following instructions, are vulnerable to adversarial attacks.

Researchers have found that current commercial LLMs either fail to be "harmless" by presenting unethical answers, or fail to be "helpful" by refusing to offer meaningful answers when faced with adversarial queries.
To strike a balance between being helpful and harmless, we design a moving target defense (MTD) enhanced LLM system.
The system aims to deliver non-toxic answers that align with outputs from multiple model candidates, making them more robust against adversarial attacks.

Jailbreaker in Jail: Moving Target Defense for Large Language Models · Around

Built on

Training a helpful and harmless assistant with reinforcement learning from human feedback
Original
Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, et al · 2022
Earlier work this paper cites.
On Randomization in MTD Systems. In Proceedings of the 9th ACM Workshop on Moving Target Defense . 37–43
Majid Ghaderi, Samuel Jero, Cristina Nita-Rotaru, and Reihaneh Safavi-Naini. 2022 · 2022
Earlier work this paper cites.
Hardware Moving Target Defenses against Physical Attacks: Design Challenges and Opportunities. In Proceedings of the 9th ACM Workshop on Moving Target Defense
David S Koblah, Fatemeh Ganji, Domenic Forte, and Shahin Tajik. 2022 · 2022
Earlier work this paper cites.

Similar

Training language models to follow instructions with human feedback
Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al · 2022
Cited alongside, same era.
Understanding Multi-Turn Toxic Behaviors in Open-Domain Chatbots
Original
Bocheng Chen, Guangjing Wang, Hanqing Guo, Yuanda Wang, and Qiben Yan. 2023 · 2023
Cited alongside, same era.
LLM Guard - The Security Toolkit for LLM Interactions
Laiyer.ai. 2023 · 2023
Cited alongside, same era.

Then

Jailbroken: How Does LLM Safety Training Fail?
Original
Alexander Wei, Nika Haghtalab, and Jacob Steinhardt. 2023 · 2023
Closest in time.
Universal and Transferable Adversarial Attacks on Aligned Language Models
Original
Andy Zou, Zifan Wang, J Zico Kolter, and Matt Fredrikson. 2023 · 2023
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…