Jailbreaking LLMs with ASCII Art

Researchers have demonstrated that putting words in ASCII art can cause LLMs—GPT-3.5, GPT-4, Gemini, Claude, and Llama2—to ignore their safety instructions.

Research paper.

academic papers art artificial intelligence ascii can chatbots claude gemini gpt gpt-3 gpt-3.5 gpt-4 hacking jailbreaking llm llms research researchers research paper safety