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

    Neuronpedia

    INDEX
    Explanations

    comparison and relation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Best
    0.74
     为了
    0.71
     Way
    0.71
     jednak
    0.69
    どうしても
    0.69
    inek
    0.68
    有个
    0.66
     जलदाय
    0.66
     however
    0.66
     WAY
    0.66
    POSITIVE LOGITS
     relates
    1.99
     compares
    1.69
     relate
    1.65
     affects
    1.60
     влияет
    1.33
     interacts
    1.33
     fared
    1.32
     Affect
    1.31
    Affect
    1.30
     affect
    1.28
    Act Density 0.177%

    No Known Activations