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

    Neuronpedia

    INDEX
    Explanations

    watchwinding accuracy

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    wning
    -0.86
     $_[
    -0.81
    aplic
    -0.79
     crushing
    -0.76
     fallecimiento
    -0.74
    itte
    -0.72
     レー
    -0.72
     tuong
    -0.71
    🥟
    -0.71
    isLogin
    -0.70
    POSITIVE LOGITS
     winding
    1.13
     accuracy
    1.07
     wound
    0.99
     seconds
    0.98
     movement
    0.98
     escape
    0.96
     Accuracy
    0.95
    winding
    0.92
    accuracy
    0.91
     movements
    0.89
    Act Density 0.018%

    No Known Activations