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

    Neuronpedia

    INDEX
    Explanations

    unconventional or unpopular

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     reliably
    0.79
     fiable
    0.71
     reliable
    0.68
     можно
    0.65
    確実に
    0.65
    aider
    0.63
     しっかり
    0.62
     สามารถ
    0.58
    健全
    0.58
     można
    0.58
    POSITIVE LOGITS
     unorthodox
    1.25
     unconventional
    1.22
     imperfect
    1.00
     unpopular
    1.00
     Controvers
    1.00
     controversial
    0.97
    Sometimes
    0.92
     Sometimes
    0.90
     Difficult
    0.88
     sometimes
    0.86
    Act Density 0.001%

    No Known Activations