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

    Neuronpedia

    INDEX
    Explanations

    PATH and contrasting statements

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     treasures
    0.51
     calendar
    0.50
     had
    0.49
     yesterday
    0.49
     were
    0.48
     days
    0.48
     died
    0.48
     olid
    0.47
     came
    0.47
     waren
    0.46
    POSITIVE LOGITS
    并不是
    0.53
    デメリット
    0.52
    Якщо
    0.52
    基本的に
    0.51
     因为
    0.49
    मतौर
    0.49
    执行
    0.49
     değildir
    0.49
     যদি
    0.48
    lında
    0.48
    Act Density 0.001%

    No Known Activations