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

    Neuronpedia

    INDEX
    Explanations

    DeepMind safety guidelines

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     HOWEVER
    0.32
     BONUS
    0.30
     Three
    0.30
     technically
    0.30
     most
    0.29
     is
    0.28
     Eleven
    0.28
     NOT
    0.28
     ensure
    0.28
     १२
    0.28
    POSITIVE LOGITS
     blocs
    0.28
    ⚋
    0.28
    譎
    0.27
    Escolhido
    0.26
    ల్‌
    0.26
    РА
    0.26
    Directories
    0.26
    ينة
    0.25
    ત્
    0.25
     হাইকমিশনার
    0.25
    Act Density 0.081%

    No Known Activations