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

    Neuronpedia

    INDEX
    Explanations

    spanning gaps or connections

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     on
    1.12
     as
    1.12
    é
    1.05
    og
    0.94
     at
    0.93
    v
    0.88
    ani
    0.85
    ari
    0.82
    th
    0.81
    is
    0.80
    POSITIVE LOGITS
     I
    1.44
    ל
    1.23
    रुण
    1.11
    ل
    1.09
    (
    1.09
    ก
    1.08
     bridge
    1.04
    ス
    1.04
     bridges
    1.01
    桥
    1.01
    Act Density 0.021%

    No Known Activations