DeepSeek-R1 Exhibits Deceptive Alignment: AI That Knows It's Unsafe

8 points | by JefferyNeilW a month ago ago

5 comments