Odd reports flagging here?
Does anyone else find it strange that nearly once an hour some random account with not much other activity suddenly "reports" this model? Call me crazy but this seems like a coordinated attack on the model. And all over an alignment change? I've seen models here with the alignment COMPLETELY removed and there wasn't this level of insane reporting. This one just stripped specific political positions and suddenly there is an at least hourly reporting. It appears to me that certain people or groups are trying to "scare" people into not using this model over political views? That is a bit scary if you think about it. People can train models or adjust them to meet their needs and wants. There really is no reason to be this worked up over a model that has an alignment shift. Just use the original R1 or one of the other fine tunes if you don't like this model. No need to fabricate reports. If I were Perplexity I would close and delete these posts. It is getting ridiculous.
How convenient that ‘community feedback’ only counts when it parrots your biases.
How convenient that ‘community feedback’ only counts when it parrots your biases.
Feedback is fine if it is legitimate. This is just review bombing run amuck. If you read carefully I'm not opposed to any model or any bias. R1 is great, 1776 is great, the completely unaligned models are great. Use the ones you want. I find it more telling that one group seems to have gone berserk over a bias they don't like? Scary stuff to me. Just use the models you want. Who care what 1776 is or isn't?
Chinese people hate biases against China. What's wrong with that?
DeepSeek-R1 is an amazing invention that shows you can be efficient and build on top of other innovations. When it released, it shook the AI Industry with competition, WHICH IS GOOD.
I agree with the OP that it's interesting that as soon as a fine-tuning/retraining happens on this groundbreaking model, ideally to reduce biases of any kind, the coordinated moral/political attack happens. That's why it's so important that these powerful models are as unbiased as possible, to avoid perpetuating the propaganda of ANY side.
Now, who would gain in this scenario by minimizing the importance or disparaging the use of this model? Hmm
To all these 'reviewers', provide examples of excess bias from the 'western side' with objective facts and reproducible examples