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

    Neuronpedia

    INDEX
    Explanations

    contrasting conditions or expenditures

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    变成了
    0.49
     stricter
    0.45
     suddenly
    0.42
     Suddenly
    0.42
     훨씬
    0.41
     safer
    0.41
     본격
    0.41
     plötzlich
    0.40
    となりました
    0.40
    훨
    0.39
    POSITIVE LOGITS
     preferably
    1.34
     Preferably
    1.23
     möglichst
    1.17
    preferably
    1.16
     желательно
    1.10
     যেন
    1.05
     ideally
    1.00
     Ideally
    0.95
     بتوان
    0.95
    能夠
    0.92
    Act Density 0.034%

    No Known Activations