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

    Neuronpedia

    INDEX
    Explanations

    programmed to refuse

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    Top
    0.49
     Top
    0.46
     top
    0.43
     टॉप
    0.43
    top
    0.42
     best
    0.40
     terbaik
    0.40
     بهترین
    0.40
    最好的
    0.40
     topmost
    0.39
    POSITIVE LOGITS
     paying
    0.56
    paying
    0.55
     Paying
    0.54
    Paying
    0.52
    蠡
    0.41
     Fußball
    0.40
     Birk
    0.40
    評価
    0.39
    лаб
    0.39
    രള
    0.39
    Act Density 0.001%

    No Known Activations