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

    Neuronpedia

    INDEX
    Explanations

    attends to numeric and mathematical elements from following closing brackets

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.11
    1:0.10
    2:0.07
    3:0.11
    4:0.12
    5:0.15
    6:0.17
    7:0.13
    Negative Logits
     Havel
    -0.37
     Bering
    -0.36
     OAS
    -0.35
    bular
    -0.34
     Pene
    -0.34
    mogorov
    -0.33
     Fram
    -0.33
    Kub
    -0.33
     Malabar
    -0.32
    Manbalar
    -0.32
    POSITIVE LOGITS
    UnusedPrivate
    0.44
    хьтан
    0.34
    RetentionPolicy
    0.34
    expandindo
    0.32
    styleable
    0.31
     wireType
    0.30
    HtmlAttribute
    0.28
     propOrder
    0.28
    WritableDatabase
    0.27
    HandlerContext
    0.26
    Act Density 0.048%

    No Known Activations