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

    Neuronpedia

    INDEX
    Explanations

    attends to various specific syntactic or formatted tokens from placeholder elements

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.20
    1:0.44
    2:0.07
    3:0.05
    4:0.05
    5:0.04
    6:0.04
    7:0.07
    Negative Logits
    /*---
    -0.38
    帖最后由
    -0.35
     Futura
    -0.34
     Avo
    -0.34
     كومونز
    -0.33
    batore
    -0.33
     ujednoznacz
    -0.33
    iosync
    -0.33
     CreateTagHelper
    -0.32
     Legisl
    -0.32
    POSITIVE LOGITS
    IsPostBack
    0.39
    expandindo
    0.35
     opérateur
    0.34
    verifyException
    0.32
    AndEndTag
    0.31
     Administrativna
    0.31
    ocidad
    0.30
     détruit
    0.30
    carpa
    0.30
     Italijani
    0.30
    Act Density 0.393%

    No Known Activations