Watermark for LLM-Generated Text

Researchers at Google have developed a watermark for LLM-generated text. The basics are pretty obvious: the LLM chooses between tokens partly based on a cryptographic key, and someone with knowledge of the key can detect those choices. What makes this hard is (1) how much text is required for the watermark to work, and (2) how robust the watermark is to post-generation editing. Google’s version looks pretty good: it’s detectable in text as small as 200 tokens.

Tags: , , , , ,

Posted on October 25, 2024 at 9:56 AM • 5 Comments

<

p id=”powered”>Sidebar photo of Bruce Schneier by Joe MacInnis.

Discover more from Cyber Elias Academy

Subscribe to get the latest posts sent to your email.

Leave a Reply

Explore More

CISA Urges Sisense Customers to Reset Credentials and Report Suspicious Activity

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) issued a warning on Thursday about a data breach at Sisense, a US business intelligence software. The agency strongly recommended that all

Read More

The World Cybercrime Index: What is it and why is it important?

Cybercriminals are notoriously tricky to pin down. They are experts in obfuscation and misdirection, masters of avoiding consequences. Not since the early days of the Wild West have criminals managed

Read More

Upcoming Speaking Engagements

Upcoming Speaking Engagements This is a current list of where and when I am scheduled to speak: I’m speaking at eCrime 2024 in Boston, Massachusetts, USA. The event runs from

Read More