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

    Neuronpedia

    INDEX
    Explanations

    twist potential

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ки
    1.05
    er
    0.96
    اب
    0.92
    ified
    0.87
    ?
    0.87
    ez
    0.86
    ,
    0.86
    ре
    0.86
    vived
    0.84
     is
    0.84
    POSITIVE LOGITS
     twist
    1.25
     twisting
    1.20
     uczni
    1.09
     twists
    1.09
     automate
    1.07
     caldo
    1.05
     imóvel
    1.05
     twisted
    1.04
    EB
    1.04
    아
    1.04
    Act Density 0.003%

    No Known Activations