INDEX
    Explanations

    The neuron detects mentions of riot‐control or crowd‐control weapons and tactics (e.g., tear gas, water cannon, rubber bullets, flash‐balls, etc.).

    New Auto-Interp
    Negative Logits
    .Http
    -0.07
     dzi
    -0.06
    -0.06
    онт
    -0.06
    σ
    -0.06
     üçüncü
    -0.06
    imizer
    -0.06
     expres
    -0.06
    ORITY
    -0.06
    (InitializedTypeInfo
    -0.06
    POSITIVE LOGITS
    -place
    0.06
     alimentos
    0.06
    ्ययन
    0.06
     atof
    0.06
    purple
    0.06
     уяв
    0.06
    dots
    0.06
     calories
    0.06
    -link
    0.06
    istorical
    0.06
    Act Density 0.008%

    No Known Activations