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

    Neuronpedia

    INDEX
    Explanations

    behavior and changes in quantities

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    军队
    0.64
    荣耀
    0.62
    公开
    0.62
     东
    0.60
    电子商务
    0.58
    联邦
    0.58
    事务
    0.57
    建设
    0.57
    情报
    0.57
    团队
    0.56
    POSITIVE LOGITS
     adsorbed
    0.65
     hysteresis
    0.62
     sinusoidal
    0.62
     spectral
    0.61
     converging
    0.61
     oscillations
    0.59
     transient
    0.59
     monotonically
    0.59
     approximately
    0.58
     gradient
    0.58
    Act Density 0.102%

    No Known Activations