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

    Neuronpedia

    INDEX
    Explanations

    detecting fraud

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    는
    1.48
    ۔
    1.45
    是
    1.34
    os
    1.30
    ه
    1.24
    a
    1.17
    是将
    1.15
    は
    1.15
    1.10
    س
    1.09
    POSITIVE LOGITS
     Fraud
    1.35
     fraud
    1.11
    Fraud
    1.09
    ;
    1.04
     defraud
    0.96
                   
    0.95
    <0x80>
    0.95
     fraude
    0.94
     fraudulent
    0.93
     SKSpriteNode
    0.89
    Act Density 0.005%

    No Known Activations