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

    Neuronpedia

    INDEX
    Explanations

    negative outcomes or difficulties

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     verhindern
    0.46
     felon
    0.44
    jdField
    0.39
     nên
    0.39
     واقعی
    0.39
    AFP
    0.39
    슨
    0.39
     xuất
    0.38
    πάρχ
    0.38
     fácil
    0.38
    POSITIVE LOGITS
    เพื่อ
    0.52
     unnecessarily
    0.52
    ne
    0.49
     unnecessary
    0.48
    k
    0.48
    為了
    0.45
     obs
    0.44
     upang
    0.43
    際に
    0.43
     pov
    0.42
    Act Density 0.001%

    No Known Activations