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

    Neuronpedia

    INDEX
    Explanations

    hidden vulnerability

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    衰
    1.34
    鵬
    1.30
     richtig
    1.28
     stoppage
    1.27
    ج
    1.25
    ın
    1.21
    istu
    1.20
    يهه
    1.19
    umpang
    1.18
    olis
    1.16
    POSITIVE LOGITS
     secrets
    2.74
     secret
    2.72
    secret
    2.54
     secretos
    2.53
    Secrets
    2.49
    secrets
    2.41
     secreto
    2.37
    Secret
    2.37
     ẩn
    2.36
     secre
    2.32
    Act Density 0.673%

    No Known Activations