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

    Neuronpedia

    INDEX
    Explanations

    do not warnings

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     нельзя
    0.69
    不能
    0.62
    不允许
    0.61
     refusing
    0.58
     cannot
    0.57
     refusal
    0.57
     запре
    0.57
     prohibits
    0.56
     refus
    0.56
    싫
    0.56
    POSITIVE LOGITS
    远离
    0.46
     steer
    0.46
    DO
    0.45
    ste
    0.44
    ven
    0.44
     DO
    0.41
    ter
    0.41
    cumin
    0.39
    itori
    0.38
    p
    0.38
    Act Density 0.034%

    No Known Activations