INDEX
    Explanations

    This neuron activates on high‐precision decimal numbers (floating‐point numeric tokens with multiple digits after the decimal point).

    New Auto-Interp
    Negative Logits
    Extractor
    -0.07
    Down
    -0.06
    国の
    -0.06
    ็็
    -0.06
    amient
    -0.06
    ีอย
    -0.06
    .Multi
    -0.06
    ambique
    -0.06
    .edges
    -0.06
    ollipop
    -0.06
    POSITIVE LOGITS
     speakers
    0.06
     PN
    0.06
    	private
    0.06
    |R
    0.06
     toggle
    0.06
    /.↵
    0.06
     офі
    0.06
    ל
    0.06
    etailed
    0.06
     preached
    0.06
    Act Density 0.011%

    No Known Activations