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

    Neuronpedia

    INDEX
    Explanations

    instead, without, unduly

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     manej
    0.33
     fyra
    0.32
     charakter
    0.31
     toutefois
    0.31
     تے
    0.31
     ਅਤੇ
    0.30
     nummer
    0.30
    いくつかの
    0.30
     sowohl
    0.30
     außerdem
    0.30
    POSITIVE LOGITS
     вместо
    0.51
     unduly
    0.46
     или
    0.45
     χωρίς
    0.45
    而不
    0.44
     unnecessarily
    0.44
     instead
    0.44
     without
    0.42
     unfairly
    0.41
     而
    0.40
    Act Density 0.121%

    No Known Activations