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

    Neuronpedia

    INDEX
    Explanations

    changing conviction

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     have
    -1.88
    IN
    -1.82
     to
    -1.77
    M
    -1.76
    ON
    -1.58
    P
    -1.55
     –
    -1.55
    OM
    -1.49
    良好的
    -1.48
    AN
    -1.47
    POSITIVE LOGITS
     about
    1.98
    〟
    1.64
     enchufe
    1.56
    TestingModule
    1.53
    以上です
    1.51
    тися
    1.50
     DECISION
    1.49
     OPINION
    1.46
    ッチリ
    1.46
    hibits
    1.45
    Act Density 0.003%

    No Known Activations