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

    Neuronpedia

    INDEX
    Explanations

    witch hunts and trials

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     ‚
    0.56
    are
    0.54
     rend
    0.54
     spesial
    0.53
    ieke
    0.51
    or
    0.50
     aly
    0.50
     bootstrapping
    0.49
     rende
    0.49
     ragion
    0.48
    POSITIVE LOGITS
     hysteria
    0.66
    י
    0.57
    Salem
    0.54
    ה
    0.53
     к
    0.51
     mêmes
    0.51
     khiển
    0.50
    Sau
    0.50
    QX
    0.49
     witches
    0.49
    Act Density 0.003%

    No Known Activations