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

    Neuronpedia

    INDEX
    Explanations

    attends to specific punctuation from related segments in parallel

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.07
    1:0.08
    2:0.07
    3:0.09
    4:0.09
    5:0.05
    6:0.38
    7:0.13
    Negative Logits
    ↵
    -0.28
    nestjs
    -0.24
    H
    -0.24
     
    -0.24
    arsal
    -0.24
    _
    -0.24
    хьтан
    -0.23
     hasta
    -0.23
    выше
    -0.23
    歹
    -0.22
    POSITIVE LOGITS
     Efq
    0.54
    NUMX
    0.50
     "..\..\
    0.48
     himſelf
    0.47
     Majefty
    0.46
    〢
    0.46
    HasForeignKey
    0.45
     cdti
    0.45
     doubtnut
    0.44
    ]='\
    0.44
    Act Density 0.046%

    No Known Activations