Jailbreaking LLMs with ASCII Art

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.

Tags: academic papers, artificial intelligence, chatbots, hacking, LLM

Posted on March 12, 2024 at 7:12 AM • 0 Comments

Leave a comment

Name

Email

URL:

Remember personal info?

Fill in the blank: the name of this blog is Schneier on ___________ (required):

Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via https://michelf.ca/projects/php-markdown/extra/

← Using LLMs to Unredact Text