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

    Neuronpedia

    INDEX
    Explanations

    making less desirable or weakening

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     understandably
    0.54
    entions
    0.42
     aufgrund
    0.41
     Laurent
    0.40
     vanwege
    0.40
     quest
    0.39
    鿆
    0.39
     explic
    0.38
     несмотря
    0.38
     explains
    0.38
    POSITIVE LOGITS
     destabil
    0.95
     disrupting
    0.87
     disrupt
    0.86
     demoral
    0.85
     discourage
    0.83
    故意
    0.81
     dissu
    0.79
     disrupts
    0.78
     disruption
    0.78
     discour
    0.77
    Act Density 0.048%

    No Known Activations