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

    Neuronpedia

    INDEX
    Explanations

    leak

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Negative Logits
    cheduling
    -0.07
     honour
    -0.06
     commanded
    -0.06
     scheduling
    -0.06
     advisors
    -0.06
     cod
    -0.06
     gross
    -0.06
    dw
    -0.06
     iterating
    -0.06
     descendant
    -0.06
    POSITIVE LOGITS
     leaks
    0.09
    -Americans
    0.08
     leak
    0.07
    ког
    0.07
    [file
    0.07
    Rain
    0.07
     Öl
    0.07
    ěn
    0.07
     shortfall
    0.07
     Sea
    0.06
    Act Density 0.005%

    No Known Activations