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

    Neuronpedia

    INDEX
    Explanations

    common

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on discourse‐structuring and emphasis markers—words that introduce, enumerate, or highlight points (e.g. “One,” “Another,” “most,” “primary,” “also,” “important”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     dikke
    -0.07
     OMAP
    -0.06
     smile
    -0.06
    Це
    -0.06
     Seat
    -0.06
     sleeves
    -0.06
     plata
    -0.06
    ゃ
    -0.06
    ρυ
    -0.06
     validators
    -0.06
    POSITIVE LOGITS
     deser
    0.07
    asurement
    0.06
    resize
    0.06
     overwrite
    0.06
    .today
    0.06
    (hours
    0.06
    ेज
    0.06
    .pres
    0.06
    -pr
    0.06
    studio
    0.06
    Act Density 2.776%

    No Known Activations