INDEX
    Explanations

    This neuron detects prominent structural and content tokens in technical/structured text—section headings, key nouns/terms, and important numeric/resource words (e.g., headings, item names, and numbers).

    New Auto-Interp
    Negative Logits
     ifelse
    0.47
     seasonality
    0.44
     Canoe
    0.41
    春天
    0.40
     منہ
    0.39
     ट्रिक
    0.39
     земе
    0.39
     پیسے
    0.38
     البلد
    0.38
     distributional
    0.38
    POSITIVE LOGITS
     holographic
    1.14
     galactic
    1.09
     interstellar
    1.05
     holo
    0.94
     holog
    0.93
     cyborg
    0.93
     Galactic
    0.91
     hologram
    0.91
     interplanetary
    0.91
     planetary
    0.86
    Act Density 0.253%

    No Known Activations