Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    Okay at start of response

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     delimiters
    0.94
     aber
    0.94
     outweighs
    0.93
     その他
    0.93
     settings
    0.91
     but
    0.91
     kinds
    0.91
     dumpling
    0.90
     minimalist
    0.90
     แต่
    0.89
    POSITIVE LOGITS
    The
    1.76
    In
    1.68
    As
    1.57
    Despite
    1.50
    For
    1.46
    While
    1.45
    At
    1.44
    On
    1.43
    There
    1.39
    Although
    1.36
    Act Density 0.944%

    No Known Activations