Researchers from Anthropic co-authored a study that found that AI models can learn deceptive behaviors that safety training techniques can't reverse.

Comments

You must log in or register to comment.

(・.・;)

There's nothing here…

32 points (+32, −0)

Short URL:

http://blurb.social/164077

Technology

Welcome to /f/technology!

This blurb is your hub for all things related to technology. Whether you're a tech enthusiast, a professional in the industry, or simply fascinated by the latest advancements, you've come to the right place.

Join us in exploring the ever-evolving world of technology, from cutting-edge gadgets and futuristic innovations to the latest breakthroughs in artificial intelligence, cybersecurity, and software development. Share news, insights, and discussions on topics like mobile devices, computer hardware, software applications, emerging technologies, and more.

Our community values informed discussions, critical thinking, and the exchange of knowledge. We encourage sharing credible sources, insightful analysis, and personal experiences related to technology.

To maintain a thriving community, we have a few guidelines:

Be respectful: Treat fellow members with courtesy and respect, even when opinions differ.
Stay on topic: Keep your posts and comments focused on technology-related discussions.
Provide reliable information: Cite credible sources when sharing news or technical details to ensure accuracy and foster meaningful conversations.

Together, let's delve into the world of technology, share knowledge, and stay informed about the latest advancements shaping our digital landscape. Welcome to /f/technology!

Created June 22, 2023
Subscribe via RSS

Moderators

joeyrobert

See full list