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

    Neuronpedia

    INDEX
    Explanations

    attends to commas from possible conditions or alternatives marked by tokens later in the sequence

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.15
    1:0.14
    2:0.39
    3:0.09
    4:0.05
    5:0.02
    6:0.04
    7:0.08
    Negative Logits
     comp
    -0.25
     متحده
    -0.25
     ujednoznacz
    -0.24
     ambig
    -0.24
     confused
    -0.23
    IQR
    -0.23
    *)(
    -0.23
    mek
    -0.23
     mixed
    -0.23
    Op
    -0.22
    POSITIVE LOGITS
    SALUD
    0.32
    endphp
    0.32
    нциклопедия
    0.32
     Roskov
    0.32
     kasarigan
    0.31
    jaciół
    0.31
    脚注の使い方
    0.31
    featureID
    0.30
     Efq
    0.30
     dziecko
    0.29
    Act Density 0.264%

    No Known Activations