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

    Neuronpedia

    INDEX
    Explanations

    excessive or negative adverbs

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     denominado
    0.99
     heterogeneity
    0.97
     hinsichtlich
    0.93
    具有
    0.88
    における
    0.85
    優れた
    0.85
     ব্যবহৃত
    0.83
     అనేక
    0.83
     полномо
    0.81
     robustness
    0.80
    POSITIVE LOGITS
     sometimes
    1.25
     horribly
    1.23
     unfairly
    1.20
     badly
    1.20
     TOO
    1.18
     uncontroll
    1.18
     terribly
    1.17
     excessively
    1.15
     awkwardly
    1.12
     harshly
    1.12
    Act Density 0.775%

    No Known Activations