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

    Neuronpedia

    INDEX
    Explanations

    text snippets

    np_max-act · gemini-2.0-flash

    comparative and superlative forms of adjectives and verbs.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron fires on runs of underscore characters (i.e. the blank “____” tokens used as placeholders in fill-in-the-blank questions).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .Exception
    -0.06
     proverb
    -0.06
    (ad
    -0.06
     ark
    -0.06
    ่ท
    -0.06
    .Brand
    -0.06
    주소
    -0.06
     Vij
    -0.06
    PRETTY
    -0.06
    -Year
    -0.06
    POSITIVE LOGITS
    izin
    0.07
    asier
    0.06
     çocu
    0.06
     نتیجه
    0.06
    ..."↵
    0.06
     YY
    0.06
    ну
    0.06
    �
    0.06
    _AX
    0.06
     quant
    0.06
    Act Density 0.005%

    No Known Activations