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

    Neuronpedia

    INDEX
    Explanations

    code constructs

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    していない
    -0.93
    duga
    -0.88
     astfel
    -0.88
    ariana
    -0.87
     tóxic
    -0.86
     Jurí
    -0.85
    瘧
    -0.84
     quella
    -0.84
    abbaye
    -0.84
     jenis
    -0.83
    POSITIVE LOGITS
    ==",
    0.95
    zación
    0.85
    culable
    0.83
    说什么
    0.82
    说出
    0.79
    Имя
    0.78
    Thal
    0.77
     увидел
    0.77
     پان
    0.77
     Thal
    0.77
    Act Density 0.004%

    No Known Activations