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

    Neuronpedia

    INDEX
    Explanations

    Best

    np_max-act · gemini-2.0-flash

    This neuron fires on “Best” (and related high-ranking list words) in titles or headings, i.e. it detects listicle-style superlatives like “Best.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     disappe
    -0.07
     Harding
    -0.06
    ;set
    -0.06
     nexus
    -0.06
    .reset
    -0.06
    ダー
    -0.06
    Par
    -0.06
    ymb
    -0.06
    idental
    -0.06
    itates
    -0.06
    POSITIVE LOGITS
     Best
    0.07
     Ś
    0.07
     Intern
    0.06
    -generator
    0.06
    ³
    0.06
     sitio
    0.06
     граждан
    0.06
     ssl
    0.06
    吸
    0.06
    UIButton
    0.06
    Act Density 0.025%

    No Known Activations