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

    Neuronpedia

    INDEX
    Explanations

    when, if, whereas, despite

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     żeby
    0.44
    哈哈
    0.41
     meestal
    0.40
    吧
    0.40
     чтобы
    0.40
     чтоб
    0.40
     patrocin
    0.40
    存档
    0.40
    证券公司
    0.40
    䜌
    0.39
    POSITIVE LOGITS
    而在
    0.61
     whereas
    0.59
     when
    0.57
     όταν
    0.56
     if
    0.55
     когато
    0.55
     quando
    0.54
     despite
    0.54
     unlike
    0.53
     wanneer
    0.53
    Act Density 0.019%

    No Known Activations