Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    all

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on the sentence-initial word “All” when it introduces a summary statement (e.g. “All groups showed…”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     recordings
    -0.07
     leak
    -0.07
    IBLE
    -0.06
     baking
    -0.06
     disturbing
    -0.06
     debtor
    -0.06
     decree
    -0.06
    _CONFIGURATION
    -0.06
     کو
    -0.06
    자료
    -0.06
    POSITIVE LOGITS
    classNames
    0.07
     sæ
    0.06
    skill
    0.06
    Pen
    0.06
     escape
    0.06
     Starts
    0.06
    (per
    0.06
     потріб
    0.06
    [this
    0.06
     arbe
    0.06
    Act Density 0.015%

    No Known Activations