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

    Neuronpedia

    INDEX
    Explanations

    would followed by verb

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     geeft
    0.53
     glad
    0.46
    4
    0.45
     люблю
    0.43
    8
    0.43
     myself
    0.42
     mulig
    0.42
     sometimes
    0.42
    侮
    0.42
     Glad
    0.41
    POSITIVE LOGITS
     customarily
    0.53
     benötigt
    0.50
     alcançar
    0.50
     Presumably
    0.47
     nécessairement
    0.47
     necessarily
    0.47
    有着
    0.47
     necessariamente
    0.46
     presumably
    0.46
     понадоби
    0.46
    Act Density 0.003%

    No Known Activations