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

    Neuronpedia

    INDEX
    Explanations

    breakdown of changes

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ا
    3.75
    ה
    3.25
    া
    3.18
    ен
    2.65
    و
    2.58
    ো
    2.56
    a
    2.50
    у
    2.46
    ة
    2.46
    jų
    2.43
    POSITIVE LOGITS
     는
    2.29
    গি
    2.19
     manslaughter
    2.15
     whistle
    2.06
     perfecting
    1.98
    apunov
    1.98
    ight
    1.96
    𝑜
    1.95
    ﺩ
    1.87
    嘌
    1.86
    Act Density 0.117%

    No Known Activations