INDEX
    Explanations

    information and details related to news articles, such as captions, images, locations, and dates

    references to mourning and memorialization in the context of events

    New Auto-Interp
    Negative Logits
    intern
    -0.70
    bid
    -0.70
    pill
    -0.69
    heit
    -0.68
    ··
    -0.66
    pillar
    -0.65
    ModLoader
    -0.65
    hops
    -0.64
    Secondly
    -0.64
    Sword
    -0.64
    POSITIVE LOGITS
     Photos
    0.91
     WATCHED
    0.80
     Caption
    0.78
     FILE
    0.78
     REUTERS
    0.77
     sidx
    0.77
     depiction
    0.76
    window
    0.75
     caption
    0.73
    ideshow
    0.72
    Act Density 0.187%

    No Known Activations