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

    Neuronpedia

    INDEX
    Explanations

    Titles and section headers

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     carnage
    0.88
     linger
    0.88
     famous
    0.84
     infamous
    0.83
     bloodshed
    0.83
     genoeg
    0.82
     middlemen
    0.82
     paraphernalia
    0.82
     thugs
    0.81
     greets
    0.80
    POSITIVE LOGITS
    Improved
    1.26
    Improving
    1.22
     Improved
    1.17
     Improving
    1.07
    Automatic
    1.04
     Automated
    1.02
     improved
    1.01
    improved
    0.98
    Enhanced
    0.97
    Autom
    0.97
    Act Density 0.190%

    No Known Activations