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

    Neuronpedia

    INDEX
    Explanations

    ID

    np_max-act · gemini-2.0-flash

    The neuron fires on numeric tokens—literal numbers (e.g. “0.609375,” hex constants) and number‐related field names/abbreviations (e.g. TOS, TTL, ID)—i.e. it detects numeric constants and numeric‐looking identifiers.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ωσε
    -0.07
     спис
    -0.07
     зазнач
    -0.07
    字段
    -0.07
    PerPixel
    -0.07
    ourced
    -0.07
    受到
    -0.06
    _CONV
    -0.06
    تو
    -0.06
    _kw
    -0.06
    POSITIVE LOGITS
    .good
    0.07
    病
    0.06
     ویژ
    0.06
    \\"
    0.06
     absolute
    0.06
    .Win
    0.06
     Communic
    0.06
    .Config
    0.06
     writers
    0.06
    .flink
    0.06
    Act Density 0.006%

    No Known Activations