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

    Neuronpedia

    INDEX
    Explanations

    socket communication

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    在
    1.16
    的山
    1.12
    しさ
    1.02
     في
    1.02
    ;
    1.01
    も
    1.00
    ﻼ
    0.97
    inine
    0.94
     в
    0.94
    رة
    0.93
    POSITIVE LOGITS
    '
    1.57
     for
    1.45
    ي
    1.42
     from
    1.31
     on
    1.27
     with
    1.24
     at
    1.24
    ع
    1.24
    at
    1.21
    т
    1.20
    Act Density 0.016%

    No Known Activations