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

    Neuronpedia

    INDEX
    Explanations

    information, panel, table entries

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    WAS
    0.99
     Is
    0.96
     Was
    0.96
     「
    0.95
     AND
    0.93
     Oed
    0.93
     IN
    0.93
     en
    0.92
    n
    0.91
     AL
    0.90
    POSITIVE LOGITS
    би
    1.01
    仕組み
    1.00
    бек
    0.99
    та
    0.94
    вого
    0.94
     подобные
    0.91
    бить
    0.91
    мена
    0.90
    choose
    0.88
    য়িক
    0.87
    Act Density 1.896%

    No Known Activations