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

    Neuronpedia

    INDEX
    Explanations

    scientific/technical writing

    np_max-act · gemini-2.0-flash

    uppercase H-initial abbreviations and identifiers, especially acronyms and code tokens like .h headers or h‑prefixed names.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    olders
    -0.07
    Пол
    -0.07
    ẳn
    -0.07
    ظة
    -0.06
     reasonable
    -0.06
     acesso
    -0.06
    olidays
    -0.06
     причин
    -0.06
     cornerback
    -0.06
    Qualifier
    -0.06
    POSITIVE LOGITS
     Harris
    0.09
     hop
    0.08
    hend
    0.08
     Harley
    0.08
     Hammond
    0.08
     Harbor
    0.08
    HL
    0.08
    -header
    0.08
     Hector
    0.08
    هار
    0.08
    Act Density 0.730%

    No Known Activations