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

    Neuronpedia

    INDEX
    Explanations

    attends to configuration-related tokens from various subsequent context tokens

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.15
    1:0.14
    2:0.30
    3:0.07
    4:0.10
    5:0.04
    6:0.06
    7:0.10
    Negative Logits
     onAnimation
    -0.28
    WriteBarrier
    -0.23
    }}^
    -0.21
     Вікіпе
    -0.21
     casado
    -0.21
    daun
    -0.21
    HNO
    -0.20
     We
    -0.20
     Our
    -0.20
    Iris
    -0.20
    POSITIVE LOGITS
     mergeFrom
    0.35
     Obrador
    0.33
     CURIAM
    0.33
    SerializedSize
    0.32
    󠁴
    0.30
     يتيمه
    0.30
    inaison
    0.30
     tịch
    0.29
     становника
    0.29
    addCriterion
    0.29
    Act Density 0.002%

    No Known Activations