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

    Neuronpedia

    INDEX
    Explanations

    randomness generation

    np_max-act · gemini-2.0-flash

    This neuron activates on tokens that signal a user’s question—especially interrogative words (“how”), pronouns around the query (“I,” “it”), and the question mark—i.e. it detects question context.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     TG
    -0.07
     Steelers
    -0.07
     X
    -0.06
     Salon
    -0.06
     막
    -0.06
     نحوه
    -0.06
    onz
    -0.06
     BMI
    -0.06
    _Font
    -0.06
    ления
    -0.06
    POSITIVE LOGITS
    fsp
    0.07
    fontName
    0.06
     SolidColorBrush
    0.06
     kıl
    0.06
    물을
    0.06
    Když
    0.06
    corner
    0.06
    .cc
    0.06
     bankrupt
    0.06
    toString
    0.06
    Act Density 0.039%

    No Known Activations