Birdwatch Note Rating
2025-02-02 14:05:43 UTC - HELPFUL
Rated by Participant: 35F7631BE8A10CC397B86144E7BA16E039D6C9266D347613F2349D192871195A
Participant Details
Original Note:
This is false and misleading. The research team recreated core functionality of a model called r1 zero which is only 3 billion parameters. Not R1 which is 671 billion. This model is significantly smaller and easier to train than R1. Read the article yourself. https://futurism.com/researchers-deepseek-even-cheaper
All Note Details