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.

Leave a Reply

Your email address will not be published. Required fields are marked *

Explore More

Friday Squid Blogging: Illex Squid and Climate Change

February 23, 2024 0 Comments 0 tags

There are correlations between the populations of the Illex Argentines squid and water temperatures. As usual, you can also use this squid post to talk about the security stories in

Businesses Increase Cybersecurity as Budgets Surge in 2024

February 22, 2024 0 Comments 0 tags

Over two-thirds of IT decision-makers increase cybersecurity budgets in 2024, prioritizing cloud security and incident response as cyber threats escalate

U-Haul Informs Customers of Major Data Breach

February 26, 2024 0 Comments 0 tags

Moving giant U-Haul has revealed that 67,000 customers were caught in a data breach last year