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

    Neuronpedia

    INDEX
    Explanations

    .com/ urls

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     but
    -2.02
     that
    -2.02
     only
    -1.95
     didn
    -1.95
    It
    -1.82
     may
    -1.80
     must
    -1.79
     should
    -1.74
    This
    -1.74
    If
    -1.70
    POSITIVE LOGITS
     soldat
    1.82
    ㈫
    1.65
     rése
    1.65
    bundet
    1.59
    gården
    1.57
    (‘
    1.50
     stadion
    1.50
     appels
    1.48
    ValueGeneration
    1.47
     gema
    1.46
    Act Density 0.069%

    No Known Activations