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

    Neuronpedia

    INDEX
    Explanations

    introduces explanation or contrast

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     thereby
    1.95
    which
    1.87
     which
    1.87
     ensuring
    1.62
     keeping
    1.61
    从而
    1.59
     allowing
    1.55
     hoping
    1.52
    Which
    1.47
     aiming
    1.44
    POSITIVE LOGITS
     exista
    0.85
     Хотя
    0.81
     quoique
    0.81
     existir
    0.80
     существует
    0.80
    Exists
    0.80
     istnieje
    0.80
     хотя
    0.80
     esiste
    0.78
     conviene
    0.75
    Act Density 0.068%

    No Known Activations