Does DeepSeek Censor Its AI Answers? On These Sensitive Topics, Yes.
-
no paywall https://archive.md/pVvYB
-
-
imgonnatrythis@sh.itjust.worksreplied to Guest 29 days ago last edited by
Well maybe yes in some aspects, but other AIs are pretty damn censored as well. At least deepseek gave me a really impressive multipage summary of pages of maths used to answer my question about how many ping pong balls can fit in the average adult vagina. Whereas Gemini just kind of shrugged me off like I was some sort of weirdo for even asking.
-
pirat@lemmy.worldreplied to Guest 29 days ago last edited by
You're asking the real important questions!
-
bell@lemmy.worldreplied to Guest 29 days ago last edited by
AI already has to deal with hallucinations, throw in government censorship too and I think it becomes even less of a serious, useful tool.
-
fuzzy_feeling@programming.devreplied to Guest 29 days ago last edited by
please let the username not check out...
-
aboubenadhem@lemmy.worldreplied to Guest 29 days ago last edited by
Making the censorship blatantly obvious while simultaneously releasing the model as open source feels a bit like malicious compliance.
-
ulrich@feddit.orgreplied to Guest 29 days ago last edited by
They're all censoring answers. They get flack every day for not censoring enough of them.
-
banshee@lemmy.worldreplied to Guest 29 days ago last edited by
Just to clarify - DeepSeek censors its hosted service. Self-hosted models aren't affected.
-
yingwu@lemmy.dbzer0.comreplied to Guest 29 days ago last edited by
I ran Qwant by Alibaba locally, and these censorship constraints were still included there. Is it not the same with DeepSeek?
-
b161@lemmy.blahaj.zonereplied to Guest 29 days ago last edited by
It just returns an error when I ask if Elon Musk is a fascist. When I ask about him generally it’s just returns propaganda.
-
banshee@lemmy.worldreplied to Guest 29 days ago last edited by
I think we might be talking about separate things. I tested with this 32B distilled model using
llama-cpp
. -
robottoaster@mander.xyzreplied to Guest 29 days ago last edited by
Even the Chinese AI is better at maths.
-
lorips@lemmy.worldreplied to Guest 29 days ago last edited by
-
tyler@programming.devreplied to Guest 29 days ago last edited by
On another person who’s actually running locally. In your opinion, is r1-32b better than Claude sonnet 3.5 or OpenAI o1? IMO it’s been quite bad, but I’ve mostly been using it for programming tasks and it really hasn’t been able to answer any of my prompts satisfactorily. If it’s working for you I’d be interested in hearing some of the topics you’ve been discussing with it.
-
lorips@lemmy.worldreplied to Guest 29 days ago last edited by
R1-32B hasn't been added to Ollama yet, the model I use is Deepseek v2, but as they're both licensed under MIT I'd assume they behave similarly. I haven't tried out OpenAI o1 or Claude yet as I'm only running models locally.
-
bytejunk@lemmy.worldreplied to Guest 29 days ago last edited by
I haven't looked into running any if these models myself so I'm not too informed, but isn't the censorship highly dependent on the training data? I assume they didn't release theirs.
-
aboubenadhem@lemmy.worldreplied to Guest 29 days ago last edited by
Video of censored answers show R1 beginning to give a valid answer, then deleting the answer and saying the question is outside its scope. That suggests the censorship isn’t in the training data but in some post-processing filter.
And even if the censorship were at the training level, the whole buzz about R1 is how cheap it is to train.
-
tyler@programming.devreplied to Guest 29 days ago last edited by
Hmm I’m using 32b from ollama, both on windows and Mac.
-
banshee@lemmy.worldreplied to Guest 28 days ago last edited by
Interesting. I wonder if model distillation affected censoring in R1.
-
psmgx@lemmy.worldreplied to Guest 28 days ago last edited by
I thought this was a big deal because it's Open Source -- is it not possible to see what is causing these blocks and censorings?
2/24