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

    Neuronpedia

    INDEX
    Explanations

    attends to punctuation marks from token pairs that involve numerical values or coordinates

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.14
    1:0.17
    2:0.14
    3:0.08
    4:0.09
    5:0.05
    6:0.09
    7:0.19
    Negative Logits
    kháu
    -0.37
    awtextra
    -0.36
    IonicModule
    -0.30
    en
    -0.26
    stateProvider
    -0.26
     Biôgrafia
    -0.25
     kræ
    -0.25
    >);
    -0.25
     Inse
    -0.24
     Walkover
    -0.24
    POSITIVE LOGITS
    tvguidetime
    0.32
     numberWith
    0.27
     doInBackground
    0.27
    PRNewswire
    0.26
    istoitu
    0.26
    VideoCapture
    0.25
    YECTO
    0.25
    viewDidLoad
    0.24
    itzender
    0.24
     Wiktionnaire
    0.24
    Act Density 0.491%

    No Known Activations