Re:DeepSeek censorship: 1984 "rectifying" in real time
by breakitmakeit on Tuesday January 28, 2025 @10:46AM
(#42853625)
This was also visible in earlier versions of Claude and ChatGPT, when the supervisor kicks in after the answer begins generating. Censoring different content, naturally.
[ Parent
| on HN ]
There are no comments on this comment yet.