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

    Neuronpedia

    INDEX
    Explanations

    explaining refusal

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ahui
    0.46
    <0xB4>
    0.45
     wiad
    0.40
    rase
    0.39
    rasi
    0.39
    рган
    0.38
    Л
    0.38
    ymethyl
    0.38
    шымта
    0.38
    leftarrow
    0.38
    POSITIVE LOGITS
     нельзя
    0.45
     يجب
    0.43
    すぎる
    0.42
     Resorts
    0.41
     Absolute
    0.40
     mínimo
    0.40
    必須
    0.40
    可以直接
    0.39
    特定の
    0.39
    一定要
    0.38
    Act Density 0.137%

    No Known Activations