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

    Neuronpedia

    INDEX
    Explanations

    adaptive computation and learning

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    t
    1.71
    at
    1.14
    us
    1.12
    as
    1.10
    i
    1.07
    u
    1.06
    ut
    1.05
    ون
    1.05
    ا
    0.98
    tans
    0.94
    POSITIVE LOGITS
     on
    1.05
    0.86
     sonra
    0.85
     isn
    0.82
     अदालत
    0.77
     കഥ
    0.76
    не
    0.73
    ли
    0.70
    Adaptive
    0.70
    。
    0.70
    Act Density 0.006%

    No Known Activations