INDEX
    Explanations

    the starts of conditional or hypothetical clauses—tokens that begin “if/imagining” style questions or hypothetical statements.

    New Auto-Interp
    Negative Logits
     Dig
    0.43
     કેટલાક
    0.40
    vdd
    0.39
    ърква
    0.38
     nigg
    0.38
     Nim
    0.37
     ஏராளமான
    0.37
    ſſ
    0.37
     Herm
    0.37
    xbet
    0.36
    POSITIVE LOGITS
     выбира
    0.75
    Choose
    0.68
     choose
    0.64
     choosing
    0.64
    choose
    0.59
     Choose
    0.59
    Choosing
    0.57
     pilih
    0.57
     memilih
    0.55
     выбрать
    0.55
    Act Density 0.074%

    No Known Activations