INDEX
    Explanations

    attends to locations related to familial relationships from relatives mentioned earlier in the sequence

    New Auto-Interp
    Head Attr Weights
    0:0.11
    1:0.14
    2:0.12
    3:0.12
    4:0.11
    5:0.03
    6:0.13
    7:0.20
    Negative Logits
    <bos>
    -0.25
    -0.24
    -0.24
     Bel
    -0.23
    eden
    -0.22
     tonel
    -0.22
     den
    -0.22
    thoscope
    -0.22
     daly
    -0.21
    leden
    -0.21
    POSITIVE LOGITS
     itſelf
    0.44
     uſed
    0.40
     ſhe
    0.40
     ſta
    0.40
     purpoſe
    0.39
     chofe
    0.39
     NSCoder
    0.39
     becauſe
    0.39
     Diſ
    0.39
     ſtate
    0.39
    Act Density 0.113%

    No Known Activations