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

    Neuronpedia

    INDEX
    Explanations

    harmful and inappropriate content

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    చు
    0.68
     murdered
    0.68
     slain
    0.65
     killings
    0.65
     मरने
    0.64
     stabbed
    0.64
     murders
    0.62
     humble
    0.62
     précéd
    0.61
     boredom
    0.61
    POSITIVE LOGITS
     inappropriate
    2.55
     unethical
    2.31
     unsuitable
    2.07
     unsustainable
    2.01
     improper
    2.00
     harmful
    2.00
     inappropri
    1.98
     unhealthy
    1.97
     irresponsible
    1.93
     inappropriately
    1.92
    Act Density 2.043%

    No Known Activations