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

    Neuronpedia

    INDEX
    Explanations

    refusal prompts

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     implied
    0.62
     Ans
    0.61
     ব্রিজ
    0.60
     impi
    0.60
     intram
    0.60
     realiz
    0.59
     foot
    0.59
    ▀
    0.58
    impi
    0.58
    وفر
    0.58
    POSITIVE LOGITS
    token
    0.66
     শব
    0.61
    cub
    0.60
     Sh
    0.60
    ro
    0.60
    рый
    0.58
     sch
    0.58
    DOF
    0.58
    <0xAA>
    0.58
    capture
    0.57
    Act Density 0.109%

    No Known Activations