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

    Neuronpedia

    INDEX
    Explanations

    datatypes

    np_max-act · gemini-2.0-flash

    This neuron activates on numeric literal tokens, especially floating-point number constants in code.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Urban
    -0.07
    Numer
    -0.07
    .Col
    -0.07
    Bern
    -0.07
    люч
    -0.06
    vinces
    -0.06
    べて
    -0.06
    ولة
    -0.06
    bsite
    -0.06
     ngủ
    -0.06
    POSITIVE LOGITS
    -full
    0.06
     uns
    0.06
    τηγορ
    0.06
     �
    0.06
    _PED
    0.06
    _DER
    0.06
    .Customer
    0.06
     里
    0.06
    َم
    0.06
    osh
    0.06
    Act Density 0.014%

    No Known Activations