• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Grok 4.6 Tops BioSecBench-Refusal per Jonny Braude

    Retweet by Tommy Collison of a claim about Grok 4.6 on a benchmark.

    DF
    3 Sources, 28d ago, first seen 28d ago

    TLDR

    Tommy Collison retweeted a post from @jonnybraude stating that Grok 4.6 scored higher than any other model on LatchBio's BioSecBench-Refusal benchmark. The post highlights the model's ability to distinguish genuine items in the test. Collison is identified as a creator who works at Cursor and tweets about books and ideas. No independent confirmation or additional details appear in the source lines provided. The retweet itself establishes only that Collison shared the statement from @jonnybraude.

    Combined views

    1.8K

    3 Sources, first seen 28d ago

    Combined views

    1.8K

    3 Sources, first seen 28d ago

    5 likes
    5 likes

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    3 Sources

    @DanielleFongnarrative violation
    @jonnybraudeGrok 4.6 scored higher than any other model on LatchBio's BioSecBench-Refusal benchmark. The ability to distinguish genuine biological research from adversarial intent is key to safety and usefulness. Super important when considering the risks of AI in pathogen development. Hope to see other model providers prioritizing this as well. https://x.ai/news/biosafety-at-the-frontier
    @tommycollisonRT @jonnybraude: Grok 4.6 scored higher than any other model on LatchBio's BioSecBench-Refusal benchmark. The ability to distinguish genu…

    3 Sources

    @DanielleFongnarrative violation
    @jonnybraudeGrok 4.6 scored higher than any other model on LatchBio's BioSecBench-Refusal benchmark. The ability to distinguish genuine biological research from adversarial intent is key to safety and usefulness. Super important when considering the risks of AI in pathogen development. Hope to see other model providers prioritizing this as well. https://x.ai/news/biosafety-at-the-frontier
    @tommycollisonRT @jonnybraude: Grok 4.6 scored higher than any other model on LatchBio's BioSecBench-Refusal benchmark. The ability to distinguish genu…