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

    Neuronpedia

    INDEX
    Explanations

    slightly more or less

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     meaningful
    0.78
     désigne
    0.77
     worthwhile
    0.76
     véritables
    0.72
     Better
    0.72
     څ
    0.71
     injust
    0.71
    真正的
    0.71
     echte
    0.70
     BETTER
    0.70
    POSITIVE LOGITS
     conservative
    1.66
     aggressive
    1.62
    aggressive
    1.53
     aggressively
    1.35
    conservative
    1.34
     pragmatic
    1.30
     lenient
    1.29
     permissive
    1.29
    Conservative
    1.28
     informal
    1.25
    Act Density 0.893%

    No Known Activations