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

    Neuronpedia

    INDEX
    Explanations

    visualizing differences and scale

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     knowing
    0.81
     оказывается
    0.73
     each
    0.72
     rou
    0.72
     found
    0.67
    intos
    0.67
     leaving
    0.67
     conduct
    0.66
     angry
    0.65
     i
    0.64
    POSITIVE LOGITS
    視覺
    1.12
     visualize
    1.07
     Visualize
    1.04
    可视化
    1.03
    Visualization
    1.03
     visualizar
    1.03
    Visualize
    1.03
    visual
    0.98
     visualiser
    0.98
     visualization
    0.95
    Act Density 0.006%

    No Known Activations