Fantastic Model
Please don't buy into the brand new accounts declaring ethical or dangerous issues with this model. They are quite obviously CCP bots angry for someone removing content that they advocate for. If your world views align with their views, feel free to use the original R1. If you prefer the alignment changes of perplexity's model then use this one. It is absolutely ok and your choice. Let's be civil here, there is plenty of room for both things to exist no? Nobody ethical review bombed the original R1, everyone celebrated it. Just relax.
Nah. It's a trashed model.
well what if ccp upload a model called llama-1949 and say that they removed the "bias" from the west? dont be double-standard.
If it's CCP bots there'll be millions of posts not hundreds and your voice will be crushed like pathetic pests, you have no clue.
well what if ccp upload a model called llama-1949 and say that they removed the "bias" from the west? dont be double-standard.
If you read what I said carefully, I was pointing out that is perfectly fine? I think there is room for many viewpoints and that is the beauty of open source. People can use reinforcement learning to shift models to suit their needs and comforts. The reaction here is pretty scary that one group is this aggressive when a model is simply re-aligned to a different cultural norm. I am saying we can simply just work with what we are comfortable with and there is literally no reason for all of this bickering? Unless someone wants their views to be the only views which again is scary and is not the way to do things.
well what if ccp upload a model called llama-1949 and say that they removed the "bias" from the west? dont be double-standard.
Of course, it's fine to have a llama-1949. Everyone can finetune a model into whatever you like, whatever idology you embrace.
Hi
If it's CCP bots there'll be millions of posts not hundreds and your voice will be crushed like pathetic pests, you have no clue.
If there were millions of posts it would generate a lot of negative attention, that can't be your goal.
Either way, you streisanded the model to the front page, so thanks for that. :*
well what if ccp upload a model called llama-1949 and say that they removed the "bias" from the west? dont be double-standard.
They have, it's called deepseek-r1. I don't see any brigading there to have it taken down, though.
well what if ccp upload a model called llama-1949 and say that they removed the "bias" from the west? dont be double-standard.
We don't care. Go upload it, knock yourself out. But you need to understand that "west" do not add any political censorship. We do not overfit our models forcing them to change 'opinions' like CCP does. That is why open-source completely uncensored models (or the models you train yourself) respond the same way as the west commercial models do about any politically sensitive topics to us. We like facts, not overfitting for political agenda. ;)
Please don't buy into the brand new accounts declaring ethical or dangerous issues with this model. They are quite obviously CCP bots angry for someone removing content that they advocate for. If your world views align with their views, feel free to use the original R1. If you prefer the alignment changes of perplexity's model then use this one. It is absolutely ok and your choice. Let's be civil here, there is plenty of room for both things to exist no? Nobody ethical review bombed the original R1, everyone celebrated it. Just relax.
Completely agree. This model rocks! 💪
I don't like that they added extra alignment to the model besides removing the CCP stuff. It's just swapping one censorship for another. Uncensored is supposed to mean uncensored.
All censorship is bad, and it doesn't matter if it's from China, USA, or HF. Knowledge belongs to the world and nobody should get to say what you're allowed to read about. The fact that censorship has become a normal part of life is absurd and I encourage everyone to fight back against ANY and ALL censorship.