INDEX
    Explanations

    drooping, open, off, empty, shrouded, false

    New Auto-Interp
    Negative Logits
     постепен
    0.68
     erfolgen
    0.64
     kształ
    0.63
    similarity
    0.63
     stabilit
    0.63
     والتي
    0.63
     chaleur
    0.62
     지속
    0.61
     त्वरित
    0.61
    安心して
    0.61
    POSITIVE LOGITS
     intact
    1.10
     tilted
    1.04
     partially
    1.03
     missing
    0.99
     locked
    0.98
     empty
    0.97
     untouched
    0.96
     turned
    0.95
     positioned
    0.94
     stuck
    0.94
    Act Density 0.657%

    No Known Activations