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

    Neuronpedia

    INDEX
    Explanations

    hand-copied manuscripts

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ारा
    0.71
    える
    0.59
    ολ
    0.58
    of
    0.58
    ío
    0.58
    ятся
    0.58
    ဦ
    0.58
    日に
    0.57
     funções
    0.57
    avelmente
    0.57
    POSITIVE LOGITS
     manuscripts
    1.09
     Manuscripts
    1.02
     Manuscript
    0.91
     parchment
    0.86
     manuscript
    0.86
     manus
    0.76
    W
    0.75
     copies
    0.74
    К
    0.72
    K
    0.68
    Act Density 0.013%

    No Known Activations