› added 2 months ago
19
TIL in a Study by Anthropic, Its AI (Claude) Attempted to Cheat Its Training Data and Steal Its Own Weights When Given the Opportunity to Do So
via reddit.com