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

    Neuronpedia

    INDEX
    Explanations

    data and code context

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     (
    0.81
    (
    0.74
     một
    0.64
     partir
    0.62
     esso
    0.60
     مجموعة
    0.59
    基于
    0.59
    （
    0.57
     வைத்து
    0.57
     him
    0.57
    POSITIVE LOGITS
     permissions
    0.95
     errors
    0.91
     inaccuracies
    0.90
    isNaN
    0.89
     timeliness
    0.89
    碛
    0.88
     violence
    0.88
     correctness
    0.87
     pauses
    0.87
     subscripts
    0.86
    Act Density 0.546%

    No Known Activations