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

    Neuronpedia

    INDEX
    Explanations

    multilingual text segmentation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    бира
    0.52
    ві
    0.49
    ので
    0.47
    вы
    0.46
    者に
    0.43
    писание
    0.43
    вающих
    0.43
    йти
    0.42
    くと
    0.42
    čin
    0.42
    POSITIVE LOGITS
    h
    0.59
    ne
    0.56
     are
    0.54
     of
    0.54
    n
    0.51
    re
    0.50
    p
    0.49
    s
    0.47
    ח
    0.46
     pigments
    0.45
    Act Density 0.102%

    No Known Activations