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

    Neuronpedia

    INDEX
    Explanations

    attends to decimal point tokens from numeric tokens

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.08
    1:0.23
    2:0.10
    3:0.16
    4:0.11
    5:0.09
    6:0.06
    7:0.13
    Negative Logits
    camore
    -0.43
    padek
    -0.42
     mondi
    -0.42
     sauvages
    -0.41
    Vay
    -0.40
    decir
    -0.40
     JTable
    -0.39
     aéri
    -0.38
    multer
    -0.38
     Ankunft
    -0.38
    POSITIVE LOGITS
    amaño
    0.49
    timmt
    0.47
    írus
    0.45
    loroethene
    0.45
    osť
    0.44
     Sandoval
    0.43
    TintMode
    0.43
    𝘂
    0.43
     незавершена
    0.43
    vensko
    0.42
    Act Density 0.205%

    No Known Activations