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

    Neuronpedia

    INDEX
    Explanations

    turns out to be true

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    setBounds
    0.47
    いずれ
    0.44
    Петербург
    0.40
    kmeans
    0.39
    ویز
    0.39
    безпе
    0.39
    ર્
    0.38
     চক্ষ
    0.38
    IY
    0.38
    வேன்
    0.38
    POSITIVE LOGITS
     ternyata
    0.89
    原來
    0.84
     surprisingly
    0.82
     actually
    0.81
     aslında
    0.80
    actually
    0.78
     Surprisingly
    0.76
    原来
    0.75
     Actually
    0.75
    果然
    0.74
    Act Density 0.140%

    No Known Activations