INDEX
    Explanations

    named entities

    This neuron detects phrases that introduce a role, function, or category (e.g. “attaccante,” “managed by,” “anchor stores”) immediately followed by a preposition (“del,” “by”) and a proper name or organization.

    New Auto-Interp
    Negative Logits
    Translated
    -0.06
    ansi
    -0.06
     MAX
    -0.06
    (op
    -0.06
     anesthesia
    -0.06
    (k
    -0.06
     Below
    -0.06
    مان
    -0.06
     containers
    -0.06
    _UDP
    -0.06
    POSITIVE LOGITS
    0.08
    ıl
    0.06
    řich
    0.06
    …"↵↵
    0.06
     svět
    0.06
    有关
    0.06
    یط
    0.06
    lášení
    0.05
    láš
    0.05
     compl
    0.05
    Act Density 0.410%

    No Known Activations