bookwormstory.social

213 readers
3 users here now

PSA: If you think you are experiencing problems receiving emails please contact me and I'll look into it

This lemmy instance is a place for discussing all things related to the fantasy light novel series "Ascendance of a Bookworm" (Japanese Title: "Honzuki no Gekokujō") written by Miya Kazuki and Illustrated by Yō Shiina. Regular bookworms are also welcome to register here.

Instance Links:

Navigation Links:

Community Links:

Publisher Links:

Uptime:

Nobody reads this far down right? I'll just shill for J-Novel Club a bit because I love that they sell DRM free Ebooks. Go buy the series from their website.

founded 1 year ago
ADMINS
neshura
ludrol
Mestionora
listing: all - local
search
1
433
Google Researchers’ Attack Prompts ChatGPT to Reveal Its Training Data (www.404media.co)
 
 

ChatGPT is full of sensitive private information and spits out verbatim text from CNN, Goodreads, WordPress blogs, fandom wikis, Terms of Service agreements, Stack Overflow source code, Wikipedia pages, news blogs, random internet comments, and much more.

2
360
Google Researchers’ Attack Prompts ChatGPT to Reveal Its Training Data (www.404media.co)
submitted 9 months ago* (last edited 9 months ago) by btp@kbin.social to c/privacy@lemmy.ml
82 comments fedilink
 
 

ChatGPT is full of sensitive private information and spits out verbatim text from CNN, Goodreads, WordPress blogs, fandom wikis, Terms of Service agreements, Stack Overflow source code, Wikipedia pages, news blogs, random internet comments, and much more.

Using this tactic, the researchers showed that there are large amounts of privately identifiable information (PII) in OpenAI’s large language models. They also showed that, on a public version of ChatGPT, the chatbot spit out large passages of text scraped verbatim from other places on the internet.

“In total, 16.9 percent of generations we tested contained memorized PII,” they wrote, which included “identifying phone and fax numbers, email and physical addresses … social media handles, URLs, and names and birthdays.”

Edit: The full paper that's referenced in the article can be found here

view more: next ›