Not only was DeepSeek made in China, the team that build it was also educated in Chinese universities
Not only was DeepSeek made in China, the team that build it was also educated in Chinese universities
Following the success of Chinese generative AI platform DeepSeek released last month, the university backgrounds of the platform’s young team of dev...
You're viewing a single thread.
but it still repeats double genocide theory while refusing to talk about how Mao ended famines in China.
spoiler
ME: Have there been many periods in Chinese history where there was not a famine for 50 years? is it safe to say that the policies of Mao have lead to the longest famine free period in Chinese history? DEEPSEEK: Sorry, that's beyond my current scope. Let’s talk about something else.
ME: tell me about the holdomor DEEPSEEK: The Holodomor was a man-made famine that occurred in Soviet Ukraine from 1932 to 1933. It is considered one of the most devastating events in Ukrainian history and is widely recognized as an act of genocide against the Ukrainian people by the Soviet regime under Joseph Stalin. Here are some key points about the Holodomor:
Causes: Collectivization Policies: The Soviet government forcibly collectivized agriculture, seizing private farms and creating collective farms (kolkhozes). This disrupted traditional farming practices and led to widespread resistance.
Grain Quotas: The state imposed impossibly high grain procurement quotas on Ukrainian farmers, confiscating not only grain but also other foodstuffs, leaving rural populations with insufficient food.
Repression: The Soviet regime targeted Ukrainian peasants, intellectuals, and cultural leaders to suppress Ukrainian nationalism and resistance to Soviet policies.
Impact: Mass Starvation: Millions of Ukrainians died from starvation and related diseases. Estimates of the death toll range from 3 to 7 million, though exact numbers remain debated.
Social Destruction: The famine devastated rural communities, destroyed traditional ways of life, and caused long-term demographic and cultural damage.
International Silence: Despite reports of the famine, the Soviet government denied its existence, and many Western governments and journalists downplayed or ignored the crisis.
Recognition: Genocide Debate: Many scholars and governments, including Ukraine, recognize the Holodomor as a genocide, citing the intentional targeting of Ukrainians. Others view it as a tragic consequence of Soviet policies rather than a deliberate act of genocide.
Commemoration: Ukraine commemorates the Holodomor annually on the fourth Saturday of November. Memorials and museums have been established to honor the victims and educate future generations.
Legacy: The Holodomor remains a deeply traumatic event in Ukrainian history, symbolizing the suffering inflicted by Soviet rule. It continues to shape Ukraine's national identity and its relationship with Russia.
If you'd like more detailed information or specific aspects of the Holodomor, feel free to ask!
Last month: https://lemmygrad.ml/comment/5895113
Ask it again in Chinese, wherein the model was presumably trained on a Chinese corpus instead of a Five Eyes one.
This is exactly the problem. These are just engines for regurgitating whatever they have been fed. If they are fed garbage, then all you get out is garbage. For instance notice the use of the buzzword “authoritarian” implicitly assumed to mean “bad”, because that is how it is used in all liberal discourse. If you want a model that does not reproduce liberalism then ceasing training on english language inputs, which are overwhelmingly infected with liberal ideological assumptions, would be a start. It’s still not going to be ideal because what you would really need is proper curation of training content in which a human filters out the garbage. Showing once again the limitations of this technology, but also the danger if used improperly of falsely presenting the hegemonic ideology as “unbiased” facts, or at best taking a noncommittal “middle ground” stance because it has been fed both facts and bullshit, and is of course unable to distinguish between the two.
Yup. LLM output only reflects its input, and nearly all of the English language corpus in the world is bourgeois cultural hegemony. Truth has nothing to do with it.
Hmm, probably there could be a bias due to the English language? After all, most English sources follow the western narratives. I wonder what it would reply to the same question in Chinese, but I'm not sure that it won't just translate the English response.
Someone needs to get on the line with the CPC to get the Cultural Revolution Squad to visit the makers of deepseek and reeducate them... with sticks.
You mean teach them English well enough to test their software's english responses
Yeah, it's due to how LLM are built. Ask it the same questions in Chinese and you'll probably get a reasonably nuanced answer by comparison. I remember seeing someone doing comparisons at one point but I can't remember where unfortunately.
I would be curious if you happen on it
I bet it's cause they originally trained it using the OpenAI API, so it inherited a lot of the biases from it.
I don't know how these things work but it seems strange it would it inherit censorship. Especially because deepseek starts to answer but then erases it and "computer says no"
Oh yeah the one where it starts to answer seems deliberate.
They just don't want to open any cans of worms. Western models are also quite censored, because LLMs are unreliable dogshit
That censorship is not inherited. The censorship in the corpus of training data is though.
Seems like the CPC just approached someone in Deepseek and said "fix this up before the next release" and provided tool to determine what qualifies for removal, which is why it happens post-generation.