INDEX
    Explanations

    words with 'snee', 'sque', or 'kne' in them, indicating a focus on onomatopoeic or sound-related terms

    New Auto-Interp
    Negative Logits
    ri
    -0.19
    sh
    -0.19
    son
    -0.19
    so
    -0.18
    sm
    -0.18
    nya
    -0.17
    hec
    -0.17
    929
    -0.16
    sg
    -0.16
    sha
    -0.16
    POSITIVE LOGITS
    aks
    0.23
    eps
    0.23
    eding
    0.21
    ez
    0.21
    eming
    0.20
    aking
    0.20
    aming
    0.20
    eper
    0.19
    ating
    0.19
    eer
    0.19
    Act Density 0.078%

    No Known Activations