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

    Neuronpedia

    INDEX
    Explanations

    gradients effectively from

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     dựng
    0.74
     sikap
    0.71
    朢
    0.70
    subplots
    0.70
     înviat
    0.67
    合わせ
    0.67
     Stand
    0.66
    Stand
    0.64
     تکن
    0.63
    ؔ
    0.62
    POSITIVE LOGITS
     flow
    3.29
     flowing
    3.23
     flows
    3.08
     flowed
    2.96
    flow
    2.84
     Flow
    2.78
    flowing
    2.69
    Flow
    2.64
     stream
    2.55
    流
    2.48
    Act Density 0.974%

    No Known Activations