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

    Neuronpedia

    INDEX
    Explanations

    reducing noise, frizz, behavior, artifacts, or interruptions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     indist
    0.39
     Depending
    0.39
     Ignoring
    0.39
     يدل
    0.39
    Ignoring
    0.38
     navigable
    0.38
     Navigation
    0.38
     mudah
    0.38
     unaltered
    0.37
    лно
    0.37
    POSITIVE LOGITS
     caused
    0.90
    caused
    0.85
     causada
    0.72
     proactively
    0.71
     altogether
    0.65
     causado
    0.64
     plag
    0.63
     CAUSED
    0.59
     efficacement
    0.58
    引起的
    0.58
    Act Density 0.051%

    No Known Activations