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

    Neuronpedia

    INDEX
    Explanations

    transparency and experience

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     perfecting
    0.46
     requires
    0.43
     welding
    0.41
     changes
    0.39
    TypeDef
    0.39
     形
    0.39
    时期
    0.38
     substandard
    0.38
     essential
    0.38
     vital
    0.38
    POSITIVE LOGITS
     transparent
    0.73
     Transparent
    0.70
     transparente
    0.67
    Transparent
    0.65
    transparent
    0.64
     прозра
    0.63
    透明
    0.59
     transpar
    0.56
    Experienced
    0.54
     Experienced
    0.54
    Act Density 0.035%

    No Known Activations