Comment by jgbuddy
5 days ago
The point of open weight models is that this is not possible, companies will be built by taking 'contraband' chinese open weight models and marketing as something from the US
5 days ago
The point of open weight models is that this is not possible, companies will be built by taking 'contraband' chinese open weight models and marketing as something from the US
A "ban" is a legal prohibition. The possibility of the ban is solely a legal decision. Open weights might make it easy to violate a ban, but it does absolutely nothing to prevent one from being possible. The questions as to whether a ban would be effective hinges entirely on the ability and desire to enforce it.
I don’t think you’re wrong but I think buddy is trying to say a ban would not be enforceable because the ability aspect won’t be there, even if the will is.
That all depends on what the law actually says and the resources and will to enforce it.
Speech/information is difficult to control, but given a draconian enough law and enforcement posture, and there can be a very significant result. You can't stop everyone, but at some point you throw enough people in prison that you've effectively suppressed it.
Not that I think that would actually happen here, but it's certainly possible.
Without a massive amount of post-training, it will be obvious that they are Chinese models with Chinese political ideology. I doubt that there's any business model that would work there.
I don't think the post-training would be that difficult. I ran an experiment on kimi k3 just now:
User: is taiwan part of china?
Kimi: Taiwan's political status is a complex and contested issue. Here's a balanced overview of the different perspectives: People's Republic of China (PRC) position: The P
<Sorry, I cannot provide this information. Please feel free to ask another question.>
I am more convinced that the Chinese models are really aligned with American values under the hood (as they likely distill US models) and the Chinese labs are the one trying to band-aid it's behavior to respond differently.
This is only if you're using the cloud version hosted in China. If you run these models locally, you don't get that
5 replies →
How are you accessing K3? If it is through Moonshot's API, then there is very likely guardrails, because it seems like it wanted to answer and was then cutoff.
We are starting to get access to K3 from US providers now, curious if they exhibit the same response pattern?
2 replies →
ask grok about the epstein files. ask gemini about the epstein files. ask them about obama's legacy as a war criminal and the reclassification of civilians as combatants to make drone striking weddings more palatable
unbelievable chauvinism in here
> Without a massive amount of post-training,
I don't think so; ISTR some LoRA thing on hugging-face that easily overrode the Tiannamen Square related weights in a previous gen GLM.
So, maybe only a few hundred dollars of training that one person does, that will "unlock" the Chinese model.
> it will be obvious that they are Chinese models with Chinese political ideology.
You aren't going to be able to prove that, not within reasonable doubt (if it's a criminal offense), nor by preponderance of evidence (if it is a civil case).
You are looking at products wrapping the popular models (i.e. moonshot, z.ai, etc) - the wrapper is doing the heavy lifting of providing guardrails. Once you have the raw array of weights and a rig with enough RAM, you can feed it subject-specific stuff to remove ideology.
do you think that China is some place devoid of business? it's the center of global commerce now. you're gravely mistaken if you think that Chinese models have "Chinese political ideology" baked into them
Grok literally had post work done to make it more right wing and racist.
sheer lunacy
Is your claim that Chinese commerce isn't subject to political censorship?
Where does Grok come into this? How does the existence of bias in one model reduce the likelihood of bias in others?
1 reply →
There are very little (read: zero) references to politics in my codebase so I don't really care if the free Chinese frontier model throws an error when I ask about Tiananmen Square. The pearl clutching about Chinese models being biased/political is a non-starter for technical work and frankly even outside of that (ie just chat capabilities, research, etc.) I think it's a little naive to think Western models aren't clearly tuned for Western bias as well. The frontier labs all have departments dedicated to "alignment" and "guardrails" that are largely driven by American political winds.
[dead]