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

    Neuronpedia

    INDEX
    Explanations

    USD Improvements

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    commitment
    0.40
    COMM
    0.39
     Commitment
    0.38
    trunk
    0.38
     بە
    0.37
    Commit
    0.37
    sour
    0.36
    MANAG
    0.36
    0.36
     ジ
    0.35
    POSITIVE LOGITS
     spec
    0.43
    Files
    0.41
     grace
    0.40
     فایل
    0.40
     pane
    0.40
     अत्यधिक
    0.39
     conex
    0.38
     client
    0.38
    Spec
    0.38
     மாறிய
    0.37
    Act Density 0.002%

    No Known Activations