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

    Neuronpedia

    INDEX
    Explanations

    slight imperfections or wear

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     explode
    0.44
     exploding
    0.44
    聋
    0.43
     rotten
    0.41
     Rotten
    0.41
    爆发
    0.39
    ujudkan
    0.38
    鸩
    0.38
    爆炸
    0.38
    รษฐกิจ
    0.37
    POSITIVE LOGITS
     imperfections
    1.02
     slight
    0.95
     blemishes
    0.87
     imperfection
    0.82
     scratches
    0.79
     Slight
    0.79
     slightly
    0.79
     faint
    0.78
     minor
    0.75
     irregularities
    0.75
    Act Density 0.022%

    No Known Activations