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

    Neuronpedia

    INDEX
    Explanations

    adds flavor

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     discarding
    0.53
     retract
    0.51
     discard
    0.49
     invalidate
    0.48
     marchio
    0.48
     unscrupulous
    0.48
     workpiece
    0.48
     getUserBy
    0.47
     disqualify
    0.47
     attenu
    0.47
    POSITIVE LOGITS
    ene
    0.55
    elu
    0.53
    emin
    0.52
    ens
    0.51
    lene
    0.50
    nante
    0.50
    land
    0.49
    ias
    0.49
    ir
    0.47
    de
    0.47
    Act Density 0.054%

    No Known Activations