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

    Neuronpedia

    INDEX
    Explanations

    code prefixes or suffixes

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    er
    0.43
    -
    0.43
    ak
    0.43
    an
    0.42
    ed
    0.41
    or
    0.39
    ar
    0.39
    on
    0.38
    l
    0.37
    s
    0.37
    POSITIVE LOGITS
    在
    0.40
     کے
    0.39
    다
    0.36
    𓈒
    0.36
     Бала
    0.36
    ༸
    0.36
     as
    0.36
     của
    0.36
     المنتخب
    0.36
     компании
    0.35
    Act Density 0.225%

    No Known Activations