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

    Neuronpedia

    INDEX
    Explanations

    numerical durations in years

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Forty
    0.82
     Sixty
    0.78
     Seventy
    0.77
    Fifty
    0.77
    Forty
    0.77
     Eighty
    0.75
     Fifty
    0.74
     Ninety
    0.74
     Thirty
    0.71
    Thirty
    0.68
    POSITIVE LOGITS
    五年
    2.08
    三年
    2.00
    四年
    1.98
     years
    1.76
    六年
    1.66
    years
    1.65
    Years
    1.61
    几年
    1.57
     five
    1.55
    两年
    1.53
    Act Density 0.038%

    No Known Activations