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

    Neuronpedia

    INDEX
    Explanations

    Transformations

    np_max-act · gemini-2.0-flash

    The neuron selectively fires on numeric tokens—especially age‐style expressions (e.g. “40 year old,” “2 year old”)—i.e. it detects mentions of numbers/ages.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    impact
    -0.07
     involved
    -0.06
     contributes
    -0.06
     �
    -0.06
    ticker
    -0.06
    요일
    -0.06
     aimed
    -0.06
     moved
    -0.06
    .map
    -0.06
    ://"
    -0.06
    POSITIVE LOGITS
     secondary
    0.07
    _GROUP
    0.06
    ');↵↵↵
    0.06
    (withId
    0.06
    賽
    0.06
     PropelException
    0.06
    >`;↵
    0.06
    .TextImageRelation
    0.06
    isSelected
    0.06
     ]↵↵
    0.06
    Act Density 0.018%

    No Known Activations