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

    Neuronpedia

    INDEX
    Explanations

    say "be"

    np_max-act-logits · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     kan
    0.78
     Kan
    0.70
    Kan
    0.67
     Can
    0.67
     Кан
    0.65
     ነበር
    0.64
    所谓的
    0.64
     trenger
    0.63
     જીવન
    0.63
    anneer
    0.63
    POSITIVE LOGITS
     be
    4.30
    be
    3.13
     Be
    3.03
    Be
    3.02
     BE
    1.95
     باشد
    1.95
     باشه
    1.86
    BE
    1.82
     soient
    1.75
     aient
    1.74
    Act Density 0.096%

    No Known Activations