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

    Neuronpedia

    INDEX
    Explanations

    model

    np_max-act · gemini-2.0-flash

    This neuron selectively activates on tokens involved in 3D-modeling instructions—especially “3”, “D”, “model”, and related creation verbs (e.g. “make”, “create”) and software names like “Blender.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    owany
    -0.07
     іс
    -0.06
    maker
    -0.06
     cel
    -0.06
    ’a
    -0.06
    ità
    -0.06
    'a
    -0.06
     slide
    -0.06
    embrance
    -0.06
     bonds
    -0.06
    POSITIVE LOGITS
    (Audio
    0.07
    _MOVE
    0.07
    ICollectionView
    0.07
     spoj
    0.06
     occasionally
    0.06
     OnInit
    0.06
     removeAll
    0.06
     withString
    0.06
     ballpark
    0.06
    .Qual
    0.06
    Act Density 0.029%

    No Known Activations