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

    Neuronpedia

    INDEX
    Explanations

    block

    np_max-act · gemini-2.0-flash

    phrases that convey quantitative or scientific information—measurements, ratios, fractions, physical processes, and spatial/relational descriptions.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    ourage
    -0.07
     harbour
    -0.07
    йтесь
    -0.07
    iệm
    -0.07
    ניות
    -0.06
    一瞬间
    -0.06
    	pthread
    -0.06
    澥
    -0.06
    ຣ
    -0.06
    аться
    -0.06
    POSITIVE LOGITS
    Middle
    0.08
        
    0.08
    _Unit
    0.07
    .char
    0.07
     usize
    0.07
     AFTER
    0.07
     Pennsylvania
    0.07
    -hop
    0.07
     HUD
    0.07
    那边
    0.07
    Act Density 8.187%

    No Known Activations