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

    Neuronpedia

    INDEX
    Explanations

    could use to ask

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     be
    1.34
     være
    1.05
     бъдат
    0.92
     happen
    0.87
     thing
    0.86
    be
    0.85
     быть
    0.85
     помочь
    0.83
     helfen
    0.83
     zostać
    0.83
    POSITIVE LOGITS
     refuses
    1.07
     advantages
    1.03
     imposes
    0.95
     avantages
    0.94
     রেখেছে
    0.94
     benefits
    0.93
     embodies
    0.92
     lacks
    0.89
     strengthens
    0.89
     prohibits
    0.89
    Act Density 0.026%

    No Known Activations