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

    Neuronpedia

    INDEX
    Explanations

    contrasting conjunctions or statements of consequence

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     vastly
    0.73
     extremely
    0.66
     European
    0.63
    非常に
    0.62
     tổng
    0.61
     tốt
    0.60
    সামরিক
    0.60
     prejudices
    0.60
     life
    0.59
     various
    0.59
    POSITIVE LOGITS
     implying
    0.77
     אך
    0.77
     Implications
    0.71
     Conversely
    0.65
    但在
    0.63
    했지만
    0.62
     suggesting
    0.62
    是因為
    0.62
     ولكن
    0.61
     immunore
    0.61
    Act Density 0.029%

    No Known Activations