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

    Neuronpedia

    INDEX
    Explanations

    names and titles

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ».
    0.50
    ிகளின்
    0.48
    》。
    0.46
    ².
    0.45
    uedata
    0.45
    ™.
    0.44
    *.
    0.44
    可能是
    0.44
    வாகும்
    0.44
    ''.
    0.44
    POSITIVE LOGITS
     chooses
    1.04
     chose
    0.94
     had
    0.90
     took
    0.90
     decides
    0.87
     insists
    0.86
     has
    0.85
     wrote
    0.85
     did
    0.84
     ने
    0.84
    Act Density 0.023%

    No Known Activations