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

    Neuronpedia

    INDEX
    Explanations

    command line arguments and code structure

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    odeon
    -0.90
    phthal
    -0.79
     enim
    -0.79
     compas
    -0.78
    わからない
    -0.78
     gd
    -0.77
    ینا
    -0.76
     preliminar
    -0.76
     pourriez
    -0.76
     jurisdic
    -0.75
    POSITIVE LOGITS
     latter
    0.96
     troisième
    0.94
    ⁝
    0.91
     gestern
    0.87
    argc
    0.82
     Whether
    0.80
     Sobald
    0.80
     third
    0.80
    みな
    0.79
    Ucraina
    0.77
    Act Density 0.014%

    No Known Activations