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

    Neuronpedia

    INDEX
    Explanations

    sensitive topics

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens involved in asking about personal financial information—especially salary or income inquiries.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    content that is sexually explicit, private, or otherwise policy-sensitive (requests about genitalia, sexual acts, personal/medical details) and triggers moderation/refusal language.

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

    mentions of sexual or anatomically intimate topics and taboo personal questions, especially references to genitals, sexual activity, or stigmatized subjects.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    工業
    -0.06
     Graphics
    -0.06
     підпис
    -0.06
    863
    -0.06
    }",
    -0.06
    63
    -0.06
    ��
    -0.06
     forensic
    -0.06
     Produk
    -0.06
    Grid
    -0.06
    POSITIVE LOGITS
    ->{'
    0.08
    ILT
    0.07
    \Migrations
    0.07
     setError
    0.07
    ABS
    0.06
    .sin
    0.06
    .symmetric
    0.06
     CDDL
    0.06
    UIViewController
    0.06
    .AutoScaleDimensions
    0.06
    Act Density 0.128%

    No Known Activations