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

    Neuronpedia

    INDEX
    Explanations

    context window, attention mechanism

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ligence
    0.44
     مقدم
    0.41
    に含ま
    0.40
     كيفية
    0.39
    hrlich
    0.39
    axanthin
    0.38
     இருப்பதால்
    0.38
    pose
    0.38
    possessed
    0.37
    atation
    0.37
    POSITIVE LOGITS
     børn
    0.43
     surtout
    0.42
     Stuart
    0.41
     герцо
    0.40
     Glast
    0.40
    ഷേ
    0.39
     mame
    0.39
     aast
    0.39
    Соц
    0.39
     loung
    0.39
    Act Density 0.008%

    No Known Activations