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

    Neuronpedia

    INDEX
    Explanations

    multi-lingual text segmentation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ation
    0.93
    ed
    0.93
    ak
    0.89
    j
    0.86
     H
    0.81
    am
    0.81
     ی
    0.80
    g
    0.78
    ched
    0.77
    k
    0.76
    POSITIVE LOGITS
    ។
    0.83
    )))
    0.82
    ם
    0.77
     ponieważ
    0.75
    ።
    0.75
    উন্ড
    0.75
    。
    0.72
    ט
    0.72
    ்ட்
    0.71
     اثنين
    0.70
    Act Density 0.072%

    No Known Activations