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

    Neuronpedia

    INDEX
    Explanations

    will

    np_max-act · gemini-2.0-flash

    This neuron activates on numeric tokens—particularly decimal numbers—such as “0.2296142578125” or other floating-point values.

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

    This neuron detects explicit sexual content, especially taboo/incestuous or coercive sexual scenarios and graphic sexual acts.

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

    explicit pornographic or fetish scenarios, especially taboo or coercive sexual content.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    /game
    -0.07
    	Grid
    -0.07
     forming
    -0.06
    -channel
    -0.06
     Naturally
    -0.06
    ひと
    -0.06
    Wrap
    -0.06
    arker
    -0.06
     glue
    -0.06
    Ted
    -0.06
    POSITIVE LOGITS
     africa
    0.07
         
    0.07
     thấy
    0.06
     край
    0.06
    .uml
    0.06
    航空
    0.06
    .addProperty
    0.06
    unal
    0.06
    +',
    0.06
     время
    0.06
    Act Density 0.495%

    No Known Activations