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

    Neuronpedia

    INDEX
    Explanations

    and then

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     undergoes
    0.46
     provides
    0.46
     describes
    0.46
     merupakan
    0.45
     constitutes
    0.45
     receives
    0.44
     indicates
    0.44
     pueden
    0.44
    वली
    0.43
     destinado
    0.43
    POSITIVE LOGITS
    尽可能
    0.57
     THEN
    0.56
    尽量
    0.55
    然后在
    0.54
    then
    0.53
     ধীরে
    0.52
    慢慢
    0.52
     Slowly
    0.51
     gradually
    0.51
     використовувати
    0.51
    Act Density 0.098%

    No Known Activations