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

    Neuronpedia

    INDEX
    Explanations

    covering topics

    np_max-act · gemini-2.0-flash

    This neuron activates on verbs and phrases that introduce or enumerate key topics or sections (e.g., “covered,” “discusses,” “include”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     workaround
    -0.08
    grading
    -0.06
     tiles
    -0.06
    	position
    -0.06
    _likelihood
    -0.06
    ering
    -0.06
     Charm
    -0.06
     Matthias
    -0.06
     Kafka
    -0.06
     billboard
    -0.06
    POSITIVE LOGITS
    icable
    0.06
    SEG
    0.06
    кар
    0.06
    ับม
    0.06
    ลอง
    0.06
    こんに
    0.06
     cn
    0.06
    bé
    0.06
    /Web
    0.06
     Rs
    0.06
    Act Density 0.055%

    No Known Activations