INDEX
    Explanations

    Japanese sentence endings

    New Auto-Interp
    Negative Logits
     seemingly
    0.92
     encompassing
    0.84
     resembling
    0.84
     potentially
    0.84
     numerous
    0.79
     evidenced
    0.78
     registered
    0.77
     spanning
    0.76
     materialized
    0.75
     two
    0.75
    POSITIVE LOGITS
    1.52
    。"
    1.27
    1.24
    ので
    1.18
    1.15
    。”
    1.11
    。」
    1.11
    。(
    1.09
    。,
    1.08
    。.
    1.08
    Act Density 0.002%

    No Known Activations