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

    Neuronpedia

    INDEX
    Explanations

    .V

    np_max-act · gemini-2.0-flash

    The neuron activates on placeholder tokens like “NAME_1,” “NAME_2,” etc., i.e. it detects when a generic NAME_x entity is mentioned.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     ips
    -0.06
     diy
    -0.06
    ERICAN
    -0.06
     Türk
    -0.06
    .definition
    -0.06
    IVE
    -0.06
    phon
    -0.06
    .red
    -0.06
    _indent
    -0.06
    ive
    -0.06
    POSITIVE LOGITS
     volatility
    0.06
     enabled
    0.06
    �
    0.06
    (Tree
    0.06
    ibilidad
    0.06
    breaker
    0.06
    adr
    0.06
     Serena
    0.06
    izzly
    0.06
    _der
    0.06
    Act Density 0.143%

    No Known Activations