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

    Neuronpedia

    INDEX
    Explanations

    introduces qualifications or emphasis

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    했는데
    0.85
    ಿದ್ದು
    0.84
    인데
    0.82
     waardoor
    0.80
     ทำให้
    0.78
     zodat
    0.76
     sehingga
    0.74
     있는데
    0.73
    。
    0.71
     waarin
    0.71
    POSITIVE LOGITS
     paradox
    1.37
     albeit
    1.34
     conversely
    1.29
     ironically
    1.29
     despite
    1.21
     unsurprisingly
    1.20
     surprisingly
    1.20
     admittedly
    1.17
    albeit
    1.15
     contrary
    1.12
    Act Density 0.300%

    No Known Activations