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

    Neuronpedia

    INDEX
    Explanations

    attends from various tokens containing names or terms to subsequent tokens that are punctuated by commas

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.07
    1:0.07
    2:0.59
    3:0.09
    4:0.05
    5:0.02
    6:0.03
    7:0.05
    Negative Logits
    RenderAtEndOf
    -0.40
    stretchr
    -0.35
     FetchType
    -0.34
    GIA
    -0.32
     Roskov
    -0.32
    BIÉN
    -0.32
      
    -0.32
     Else
    -0.31
    endfor
    -0.31
    testcase
    -0.31
    POSITIVE LOGITS
     Normdatei
    0.40
    homonymie
    0.35
     alimentaria
    0.32
    javac
    0.32
    verwijspagina
    0.32
     يتيمه
    0.32
    bootstrapcdn
    0.31
    usercontent
    0.31
    Fabian
    0.31
     Raccoon
    0.31
    Act Density 0.340%

    No Known Activations