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

    Neuronpedia

    INDEX
    Explanations

    attends to various tokens, often signifying a change or command, from other tokens that identify or specify the context or category

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.09
    1:0.11
    2:0.10
    3:0.08
    4:0.08
    5:0.02
    6:0.15
    7:0.32
    Negative Logits
     AssemblyVersion
    -0.41
    Tazama
    -0.37
    Atentamente
    -0.35
     mxArray
    -0.34
     ***!
    -0.33
    férences
    -0.33
    azia
    -0.32
    fromnode
    -0.32
     للاسماء
    -0.32
    }],
    -0.31
    POSITIVE LOGITS
    期刊论文
    0.35
    bilt
    0.33
     Mard
    0.31
    OnPage
    0.31
     referenties
    0.31
    𝘤
    0.31
     Vanderbilt
    0.30
     Minato
    0.30
     bege
    0.30
    osoba
    0.30
    Act Density 1.169%

    No Known Activations