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

    Neuronpedia

    INDEX
    Explanations

    attends to numeric tokens from themselves and adjacent tokens

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.07
    1:0.21
    2:0.09
    3:0.04
    4:0.15
    5:0.25
    6:0.07
    7:0.08
    Negative Logits
     للمعارف
    -0.40
     للاسماء
    -0.38
     дописавши
    -0.37
     linkovi
    -0.33
    ValueGeneration
    -0.32
    FontOfSize
    -0.31
     nakalista
    -0.31
    ConstraintMaker
    -0.31
     joaat
    -0.28
     Italijani
    -0.28
    POSITIVE LOGITS
    unately
    0.30
    XmlAccessType
    0.29
    outheast
    0.27
    skrä
    0.26
    edly
    0.26
    eterangan
    0.26
     cioc
    0.26
    martre
    0.26
     anzu
    0.25
    OMET
    0.25
    Act Density 0.104%

    No Known Activations