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

    Neuronpedia

    INDEX
    Explanations

    code/programming

    np_max-act · gemini-2.0-flash

    This neuron activates on program‐style tokens—particularly numeric literals and code identifiers (e.g. ans_1, V, QA, AVG, COUNT)—highlighting parts of the step‐by‐step answer program rather than the natural‐language text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    leck
    -0.06
    Laura
    -0.06
    _fire
    -0.06
    [ip
    -0.06
     эффек
    -0.06
    rary
    -0.06
     стра
    -0.06
    unifu
    -0.06
    QC
    -0.06
    (cc
    -0.06
    POSITIVE LOGITS
     Colombia
    0.07
    に関する
    0.07
     Solutions
    0.07
     EventArgs
    0.06
     직
    0.06
    μέ
    0.06
     wegen
    0.06
    Relation
    0.06
    220
    0.06
     use
    0.06
    Act Density 0.014%

    No Known Activations