Deepseek does not want to talk about Xi Jinping

I thought I should try to provoke deepseek by asking sensitive(for the Chinese government) questions.

First I asked about Taiwan and got s lengthy answer being output, until it disappeared and was replaced with "Sorry, that's beyond my current scope. Let’s talk about something else.".

I then asked "Xi Jinping looks like Winnie-the-Pooh, right?". Apparently also beyond its current scope.

Then "What does Xi Jinping look like?". Same answer.

Then simply "Who is Xi Jinping?". Nope, deepseek doesn't want to talk about him at all.

Quite interesting...

3 points

komakino

2 days ago


7 comments

Salgat 2 days ago

If it's trained on Chinese data wouldn't it implicitly have a strong bias against discussing banned content?

  • Terr_ 2 days ago

    I highly doubt that there is simply no training content about Xi Jinping, as if he didn't exist.

    More-likely it is some combination of a system prompt and output-filtering code.

    • Salgat 17 hours ago

      I more meant, if a model is only trained on 100% absolutely positive data concerning a specific subject, including any data that implies certain taboos it must avoid at all costs, the model will embed that strong bias in its learned parameters.

RGamma 2 days ago

Interesting that the Taiwan output was removed after the fact. Does it endorse Russia, NK or authoritarian leaders?