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

    Neuronpedia

    INDEX
    Explanations

    declining harmful requests

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     have
    0.65
     are
    0.61
     Pode
    0.61
     for
    0.60
    蘄
    0.58
    ط
    0.57
     Como
    0.57
    8
    0.57
     şi
    0.56
    for
    0.55
    POSITIVE LOGITS
    an
    0.65
    ่
    0.62
    д
    0.61
    ↵
    0.55
    м
    0.54
    ان
    0.52
    िकता
    0.50
    া
    0.50
    ELINE
    0.50
     cannabin
    0.49
    Act Density 0.063%

    No Known Activations