© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. MLP Neurons - Layer 5
    7. 9092
    Prev
    Next
    Layer 4Layer 6
    INDEX
    Explanations

    osa (in Espinosa)

    np_max-act · claude-sonnet-5-5
    Triggered by @paultiplady

    osa

    np_max-act-logits · claude-sonnet-5-5
    Triggered by @paultiplady

    spinach Method used: 1, because the max activating tokens are pieces of the word "spinach" in various languages (菠菜, espinacas, spinach, Spinaci)

    np_min-act-logits · claude-sonnet-5-5Neg Acts
    Triggered by @paultiplady

    the "osa" token completing the surname "Espinosa" (strongest on "osa" after "Esp" + "in"), with weaker activation on the "ino"/"in" pieces of related name fragments.

    oai_token-act-pair · claude-sonnet-5-5
    Triggered by @paultiplady

    The final subword token "osa" completing the surname "Espinosa" (after the "Espin" fragment), across many different people and contexts. It also fires, more weakly, on a similar name-final "a" in "Robirosa".

    eleuther_acts_top20 · claude-sonnet-5-5
    Triggered by @paultiplady
    New Auto-Interp
    Top Features by Cosine Similarity
    Embeds
    IFrame
    Link
    Loading dashboard...