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

    Neuronpedia

    INDEX
    Explanations

    first choice or step

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     thing
    0.61
     Thing
    0.53
    thing
    0.48
     coisa
    0.48
    Thing
    0.48
     şeyi
    0.43
     cosa
    0.43
     coisas
    0.41
     ነገር
    0.41
     primeiras
    0.41
    POSITIVE LOGITS
     choice
    0.68
     expérience
    0.59
     choix
    0.58
     keuze
    0.57
     experience
    0.56
    expérience
    0.55
    choice
    0.54
     instinct
    0.54
    經驗
    0.54
     experiencia
    0.52
    Act Density 0.009%

    No Known Activations