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

    Neuronpedia

    INDEX
    Explanations

    cannot provide or fulfill

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     meestal
    0.55
     mutlaka
    0.48
     Wrapping
    0.47
     meist
    0.46
     biraz
    0.46
     sempre
    0.45
     Ensuring
    0.45
     wrapping
    0.45
     helemaal
    0.45
    其余
    0.44
    POSITIVE LOGITS
     ever
    1.05
     EVER
    1.03
     knowingly
    0.92
     engage
    0.89
    ever
    0.78
     cond
    0.74
     deviate
    0.74
     jeopardize
    0.74
    engage
    0.73
     interfere
    0.72
    Act Density 0.186%

    No Known Activations