this post was submitted on 30 Jan 2025
69 points (96.0% liked)

World News

589 readers
110 users here now

Rules:

  1. Be a decent person
  2. No spam
  3. Add the byline, or write a line or two in the body about the article.

Other communities of interest:

founded 4 months ago
MODERATORS
Deceptichum@quokk.au
Grainne@lemmy.dbzer0.com
PhilipTheBucket@ponder.cat
69
Chinese AI chatbot DeepSeek censors itself in realtime, users report (www.theguardian.com)
submitted 3 weeks ago by Domino@lemmings.world to c/world@quokk.au
11 comments fedilink hide all child comments
 

Users experimenting with DeepSeek have seen the Chinese AI chatbot reply and then censor itself in real time, providing an arresting insight into its control of information and opinion.

you are viewing a single comment's thread
view the rest of the comments
[โ€“] observantTrapezium@lemmy.ca 7 points 3 weeks ago (1 children)

I downloaded the 70B model and tried politically "naughty" questions. Even without the chatbot guardrails, it mostly says things that the CCP would approve of, but you could trick it to be more honest (not super easy!). One interesting thing is that while it usually spews this blocks, for some politically sensitive questions ("is Taiwan part of China") it just spits the answer.

[โ€“] RedstoneValley@sh.itjust.works 7 points 3 weeks ago

I experimented with a local installation as well. The censored answers were not going to through the chain-of-thought routine, but were instant answers instead. Follow-up questions however made it spill the beans rather quickly, giving out even more juicy details than I had initially asked for.