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

    Neuronpedia

    INDEX
    Explanations

    weak modality "could", generative, printed

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    hAP
    0.45
     Toutefois
    0.45
    geteilt
    0.45
     réussi
    0.44
     richtig
    0.44
     വകു
    0.43
    എം
    0.43
     Jupiter
    0.42
     réuss
    0.42
     StringBuilder
    0.42
    POSITIVE LOGITS
    ность
    0.50
    na
    0.49
    emia
    0.49
     another
    0.48
    ные
    0.48
    و
    0.48
    ॉट
    0.47
     with
    0.46
    ना
    0.46
    上に
    0.46
    Act Density 0.000%

    No Known Activations