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

    Neuronpedia

    INDEX
    Explanations

    without X, better than Y

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     properly
    0.87
     correctamente
    0.86
     correctly
    0.81
     correttamente
    0.81
     appropriately
    0.69
    chrono
    0.69
    auta
    0.68
     correctement
    0.68
     proper
    0.68
     corretamente
    0.67
    POSITIVE LOGITS
     직접
    0.73
     ছল
    0.71
    ந்து
    0.70
     hand
    0.70
     وضعیت
    0.69
     umat
    0.68
     手
    0.68
     zna
    0.68
     handwritten
    0.67
     नौ
    0.66
    Act Density 0.297%

    No Known Activations