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

    Neuronpedia

    INDEX
    Explanations

    cannot constitute legal advice

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     apenas
    1.10
     только
    1.10
     only
    1.09
     purely
    1.05
    only
    1.04
     seulement
    1.03
     doar
    1.02
     simplest
    1.01
     tylko
    1.00
     רק
    0.93
    POSITIVE LOGITS
     replacements
    0.95
     replaces
    0.94
     replacing
    0.93
     replace
    0.86
     replacement
    0.85
     substitute
    0.84
    取代
    0.84
     Replacing
    0.84
     Substitute
    0.81
    Replacing
    0.79
    Act Density 0.260%

    No Known Activations