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

    Neuronpedia

    INDEX
    Explanations

    attends to numerical tokens from other similar numerical tokens listed earlier in the sequence

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.13
    1:0.14
    2:0.14
    3:0.12
    4:0.12
    5:0.03
    6:0.12
    7:0.17
    Negative Logits
     Rom
    -0.26
    CCN
    -0.24
     Reg
    -0.24
    слава
    -0.23
     Martens
    -0.23
    obraz
    -0.23
    <h1>
    -0.23
    ску
    -0.23
     rég
    -0.22
     Revolution
    -0.21
    POSITIVE LOGITS
    ſelf
    0.48
     متعلقه
    0.47
     فريبيس
    0.46
    LookAnd
    0.45
     resourceCulture
    0.45
     myſelf
    0.43
     himſelf
    0.43
    Искәрмәләр
    0.42
     betweenstory
    0.41
    IUrlHelper
    0.41
    Act Density 0.166%

    No Known Activations