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

    Neuronpedia

    INDEX
    Explanations

    difficult to memorize

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     reluctantly
    0.89
     unbearable
    0.88
     uncomfortable
    0.84
     unacceptable
    0.82
     unusable
    0.82
     poor
    0.81
     rough
    0.81
     unnecessary
    0.81
     unwillingness
    0.80
    緊
    0.80
    POSITIVE LOGITS
     accidentally
    1.24
     easily
    1.14
     leakage
    1.11
     leaks
    1.10
     mistakes
    1.08
     accidents
    1.02
    Errors
    1.02
     Easily
    1.02
     بسه
    1.01
     theft
    1.00
    Act Density 0.509%

    No Known Activations