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

    Neuronpedia

    INDEX
    Explanations

    quantization and precision

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     hardly
    1.29
    不高
    1.22
    这一
    1.18
     distintas
    1.15
     একথা
    1.03
     distinto
    1.03
     stiffness
    1.02
    WOM
    1.01
     impatiently
    1.00
     prevenção
    1.00
    POSITIVE LOGITS
    ……….
    1.21
    ………..
    1.19
    öglichkeiten
    1.16
    ……..
    1.15
    і
    1.14
    ẖ
    1.12
    рте
    1.11
     heme
    1.11
    การ
    1.10
    韫
    1.10
    Act Density 0.001%

    No Known Activations