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

    Neuronpedia

    INDEX
    Explanations

    server verifies

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     thought
    0.82
     think
    0.82
     heavily
    0.80
     clumsy
    0.77
     DIY
    0.77
     thinks
    0.76
     sneaky
    0.75
     kreativ
    0.75
     really
    0.75
     thinking
    0.75
    POSITIVE LOGITS
     Review
    1.02
     review
    1.01
     reconcile
    1.01
    Review
    0.98
    review
    0.97
     notify
    0.97
    Compare
    0.96
     notifies
    0.94
     verifies
    0.93
     Reviewed
    0.93
    Act Density 0.245%

    No Known Activations