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

    Neuronpedia

    INDEX
    Explanations

    Select operation, Choose operation, Decoding, Remove, merging

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    0
    0.41
    ,"
    0.41
    <start_of_image>
    0.41
    ,”
    0.40
     cr
    0.39
    Cr
    0.39
    cr
    0.37
     good
    0.36
    ."
    0.36
    —”
    0.36
    POSITIVE LOGITS
    䢔
    0.61
    㪇
    0.57
     inhibitory
    0.55
    逪
    0.54
    ရိ
    0.54
    晎
    0.52
    䘚
    0.52
    䢍
    0.52
    唕
    0.52
    㣱
    0.52
    Act Density 0.001%

    No Known Activations