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

    Neuronpedia

    INDEX
    Explanations

    harmful instructions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     ambush
    0.43
     vegetal
    0.40
     ஏன்
    0.39
    дни
    0.38
     kut
    0.37
    廼
    0.37
     woord
    0.36
     amen
    0.36
    Архівовано
    0.35
    ദാ
    0.35
    POSITIVE LOGITS
    alte
    0.36
    Brit
    0.35
    Ext
    0.35
    Cornell
    0.35
     সোভ
    0.34
    ពួកគេ
    0.34
    履行
    0.33
     Britannica
    0.33
     López
    0.33
     ग्रेटर
    0.33
    Act Density 0.001%

    No Known Activations