Skip to main content

Uncensored Deepseek R1 671B

Already open-source and available on hugging face, Deepseek R1 671B fine tuned by Perplexity, dubbed “perplexity-ai/r1-1776” on hugging face.

Status: Completed2 comments

Log in to comment and vote

Comments2

  • Silver River

    •

    Feb 23, 2025

    Definitely, I’ve literally asked a question about a company and the current model tried to enforce it’s employee guidelines on it’s response when it has no reason to follow them! And I’m not talking about Venice obviously, in it’s current state DeepSeek R1 671B on Venice is hard to find a use case for when there’s Llama 3.1 405B because of it’s behaviours imo.

  • JP

    Team•

    Feb 23, 2025

    We are evaluating this model, but do not have a firm timeline for if it will be rolled out.