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

    Neuronpedia

    INDEX
    Explanations

    shapes and dimensions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     for
    0.80
    in
    0.57
     Operations
    0.54
    for
    0.53
     Over
    0.52
     Programs
    0.51
     Nodes
    0.51
     Solutions
    0.50
     Tes
    0.50
     Transforms
    0.50
    POSITIVE LOGITS
    :
    0.68
    שה
    0.55
    อย่า
    0.55
    щены
    0.52
    是
    0.51
    agonal
    0.50
     concatenated
    0.48
    不要
    0.48
    ה
    0.48
    Initialized
    0.48
    Act Density 0.010%

    No Known Activations