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

    Neuronpedia

    INDEX
    Explanations

    hypothetical wishes and questions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    非常に
    0.92
     фактически
    0.91
    Subsequently
    0.88
    试图
    0.86
     subsequently
    0.85
     subsequent
    0.81
     presumably
    0.81
    highest
    0.81
     następnie
    0.78
     تقریبا
    0.78
    POSITIVE LOGITS
     rewind
    1.19
     magic
    1.11
     magical
    1.04
     mágico
    0.95
     timpul
    0.92
     puedo
    0.91
     magia
    0.88
     alchemy
    0.87
     mág
    0.85
     glimpse
    0.84
    Act Density 0.021%

    No Known Activations