INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Negative Logits
     erroneous
    0.81
     misleading
    0.75
     differed
    0.74
     flimsy
    0.72
     negative
    0.69
     withstood
    0.68
     required
    0.67
     vantage
    0.67
     included
    0.66
     specified
    0.66
    POSITIVE LOGITS
    が好き
    1.23
     люблю
    1.21
     hobbies
    1.10
     любит
    1.08
    Hobbies
    1.01
    爱好
    0.96
     passionately
    0.96
    哲学
    0.96
    喜歡
    0.95
     sonhos
    0.95
    Act Density 0.005%

    No Known Activations