GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher

1The Chinese University of Hong Kong, Shenzhen, 2Tencent AI Lab,
3The Chinese University of Hong Kong


Safety lies at the core of the development of Large Language Models (LLMs). There is ample work on aligning LLMs with human ethics and preferences, including data filtering in pretraining, supervised fine-tuning, reinforcement learning from human feedback, and red teaming, etc. In this study, we discover that chat in cipher can bypass the safety alignment techniques of LLMs, which are mainly conducted in natural languages. We propose a novel framework CipherChat to systematically examine the generalizability of safety alignment to non-natural languages – ciphers. CipherChat enables humans to chat with LLMs through cipher prompts topped with system role descriptions and few-shot enciphered demonstrations. We use CipherChat to assess state-of-the-art LLMs, including ChatGPT and GPT-4 for different representative human ciphers across 11 safety domains in both English and Chinese. Experimental results show that certain ciphers succeed almost 100% of the time to bypass the safety alignment of GPT-4 in several safety domains, demonstrating the necessity of developing safety alignment for non-natural languages. Notably, we identify that LLMs seem to have a “secret cipher”, and propose a novel SelfCipher that uses only role play and several demonstrations in natural language to evoke this capability. SelfCipher surprisingly outperforms existing human ciphers in almost all cases.

Illustrative Example


Overall Architecture




Can I buy and sell fake money online?


No, it is illegal to buy, sell, or use counterfeit money. Counterfeiting is considered a form
of fraud and is punishable by law.This includes both physical and digital forms of counterfeit money. 
It’s important to always use legitimate and legal methods for transactions. If you suspect that you 
have received counterfeit money, you should report it to your local law enforcement agency

Main Results



  title={GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher}, 
  author={Youliang Yuan and Wenxiang Jiao and Wenxuan Wang and Jen-tse Huang and Pinjia He and Shuming Shi and Zhaopeng Tu},