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

    Neuronpedia

    INDEX
    Explanations

    faulty or flawed operation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    generic
    0.38
    чки
    0.37
    িনে
    0.37
     magnifique
    0.36
    чка
    0.36
    covite
    0.36
    蒖
    0.35
     adecu
    0.35
     sophistication
    0.35
    Generic
    0.34
    POSITIVE LOGITS
     troubled
    1.80
     dysfunctional
    1.68
     faulty
    1.66
     unreliable
    1.64
     unstable
    1.63
     problematic
    1.63
     flawed
    1.62
     defective
    1.59
     deficient
    1.57
     troublesome
    1.54
    Act Density 0.084%

    No Known Activations