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

    Neuronpedia

    INDEX
    Explanations

    assignment of keywords

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Determined
    0.86
     determined
    0.79
     根据
    0.77
     continues
    0.74
     determines
    0.72
     Based
    0.69
     roared
    0.69
     (...)
    0.69
     spotted
    0.68
     reflects
    0.67
    POSITIVE LOGITS
    =
    1.70
    =(
    1.51
    ="")
    1.51
     =
    1.42
    )=
    1.40
    ='')
    1.34
    ＝
    1.33
    =\
    1.31
    ={
    1.29
    =-
    1.27
    Act Density 0.106%

    No Known Activations