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

    Neuronpedia

    INDEX
    Explanations

    too sensitive / a bit big

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     markedly
    0.90
     inherently
    0.88
     engender
    0.85
     readily
    0.83
     appreciably
    0.82
     ostensibly
    0.82
    颇
    0.81
     propensity
    0.80
     intrinsically
    0.80
     predominantly
    0.79
    POSITIVE LOGITS
     weird
    1.91
     scary
    1.89
     annoying
    1.84
     funny
    1.83
     creepy
    1.72
    weird
    1.72
     silly
    1.71
     boring
    1.69
     crazy
    1.61
    funny
    1.61
    Act Density 0.445%

    No Known Activations