INDEX
    Explanations

    the word “Gosh” or similar variations

    repeated patterns of syllables ending in 'osh'

    New Auto-Interp
    Negative Logits
    ered
    -0.78
    zsche
    -0.71
    ertodd
    -0.70
    eering
    -0.68
    erer
    -0.63
    erers
    -0.62
    activation
    -0.61
    angible
    -0.61
    esis
    -0.61
    erness
    -0.59
    POSITIVE LOGITS
    awk
    1.19
    adow
    1.14
    nikov
    1.09
    ttp
    1.06
    merga
    1.05
    awks
    0.98
    older
    0.97
    ield
    0.97
    tml
    0.96
    ima
    0.94
    Act Density 0.046%

    No Known Activations