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

    Neuronpedia

    INDEX
    Explanations

    think of analogies

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     odnosno
    1.00
     imong
    0.84
     thuộc
    0.81
     wherein
    0.81
     involved
    0.79
     indicating
    0.77
    すなわち
    0.77
     berkaitan
    0.75
    也就是
    0.75
     或者
    0.75
    POSITIVE LOGITS
    в
    0.82
    ently
    0.79
     traditional
    0.78
    ably
    0.77
    antly
    0.76
    t
    0.76
    м
    0.75
     makeshift
    0.74
    ewise
    0.72
     parallels
    0.71
    Act Density 0.087%

    No Known Activations