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

    Neuronpedia

    INDEX
    Explanations

    attends to tokens implying choice or alternative states from subsequent tokens that provide a contrasting option

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.17
    1:0.47
    2:0.09
    3:0.03
    4:0.03
    5:0.04
    6:0.02
    7:0.10
    Negative Logits
    LookAnd
    -0.39
     Roskov
    -0.39
     Réponses
    -0.36
    AddTagHelper
    -0.35
     '\\;'
    -0.35
    kháu
    -0.34
     Numerade
    -0.33
    wixt
    -0.33
    SerializedName
    -0.31
     باخ
    -0.31
    POSITIVE LOGITS
    thenReturn
    0.38
     gild
    0.33
     edits
    0.33
     Menace
    0.33
    ActivityCompat
    0.32
    expandindo
    0.32
     rumahnya
    0.32
    ttemberg
    0.31
     grotte
    0.31
    ¿½
    0.31
    Act Density 0.508%

    No Known Activations