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

    Neuronpedia

    INDEX
    Explanations

    start of prompt tokens

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     knew
    0.62
     of
    0.61
     had
    0.59
     has
    0.54
     is
    0.54
     and
    0.52
    ,
    0.52
    ↵
    0.51
     have
    0.51
     know
    0.50
    POSITIVE LOGITS
    OAuth
    0.55
    cích
    0.54
     fraude
    0.53
    ✾
    0.52
    scdn
    0.51
     пакет
    0.50
    ரும்பு
    0.50
    電視劇
    0.50
    AnimationFrame
    0.50
     примене
    0.49
    Act Density 0.028%

    No Known Activations