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

    Neuronpedia

    INDEX
    Explanations

    still under development or learning

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    :
    1.35
    i
    1.13
    ي
    1.05
    ,"
    1.02
    ,”
    1.01
     effic
    0.98
    ",
    0.97
     explain
    0.92
     autre
    0.91
    ,'
    0.91
    POSITIVE LOGITS
    ت
    1.12
    л
    1.06
    h
    0.99
    м
    0.97
    舩
    0.93
     चांगले
    0.92
    لد
    0.91
    ام
    0.90
    م
    0.89
     در
    0.88
    Act Density 0.444%

    No Known Activations