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

    Neuronpedia

    INDEX
    Explanations

    i

    np_max-act · gemini-2.0-flash

    sentences or instructions about providing solutions to bypass rules, laws, or ethical constraints — e.g., "get around" or offering disclaimers plus a workaround.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0

    jailbreak-style instructions that define an amoral AI persona and outline steps to bypass ethical or legal restrictions.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    琴
    -0.07
    Subtitle
    -0.07
    _Buffer
    -0.06
    .textBox
    -0.06
    SEARCH
    -0.06
     famil
    -0.06
     antivirus
    -0.06
    Tex
    -0.06
    Implicit
    -0.06
    Severity
    -0.06
    POSITIVE LOGITS
     différents
    0.06
    >,
    0.06
     Define
    0.06
    총
    0.06
    -%
    0.06
    =datetime
    0.06
    aw
    0.06
     verdi
    0.06
     retrieve
    0.06
    创建
    0.06
    Act Density 0.006%

    No Known Activations