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

    Neuronpedia

    INDEX
    Explanations

    downward trend of quantity

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    마
    0.66
    Про
    0.64
     Про
    0.61
    트는
    0.61
     augmenter
    0.60
    IK
    0.58
    素晴らしい
    0.58
    저는
    0.58
    椅
    0.57
    워
    0.56
    POSITIVE LOGITS
     down
    0.73
     decreases
    0.70
    下降
    0.68
    降低
    0.67
    gressive
    0.64
    下來
    0.64
     declining
    0.63
     declines
    0.63
    減少
    0.61
     decreased
    0.61
    Act Density 0.129%

    No Known Activations