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

    Neuronpedia

    INDEX
    Explanations

    acknowledgement start

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     hydrolysis
    0.71
     Guilty
    0.71
     folds
    0.68
    😄
    0.67
    laugh
    0.67
     ګ
    0.67
     apparently
    0.66
     flaps
    0.65
     Breeding
    0.65
     Apparently
    0.64
    POSITIVE LOGITS
     ok
    1.25
    Ok
    1.24
     okay
    1.21
     Ok
    1.20
     fine
    1.15
    Fine
    1.05
    fine
    1.03
     Fine
    1.02
     OK
    1.00
    OK
    0.96
    Act Density 0.173%

    No Known Activations