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

    Neuronpedia

    INDEX
    Explanations

    Mid-Range

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    IZER
    0.84
     analyzes
    0.81
     neighbors
    0.81
    izations
    0.80
    IZATION
    0.80
     defense
    0.80
    izin
    0.80
     violation
    0.79
    izes
    0.78
     initialization
    0.78
    POSITIVE LOGITS
     whilst
    0.99
    organised
    0.97
    Whilst
    0.97
     Whilst
    0.94
     فى
    0.93
     spoilt
    0.88
     organised
    0.87
     bespoke
    0.82
     (£
    0.80
    プレゼ
    0.79
    Act Density 0.015%

    No Known Activations