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

    Neuronpedia

    INDEX
    Explanations

    code/logs/internet messages

    np_max-act · gemini-2.0-flash

    This neuron specifically detects the special “<|end_header_id|>” token.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    The neuron detects special protocol/control tokens that mark headers or other metadata boundaries in the input token stream.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
    do
    -0.07
    Він
    -0.07
    InRange
    -0.07
    Two
    -0.07
    five
    -0.07
    easy
    -0.06
     후기
    -0.06
    their
    -0.06
    дать
    -0.06
     five
    -0.06
    POSITIVE LOGITS
     INTER
    0.06
     reck
    0.06
    .onreadystatechange
    0.06
    �
    0.06
     RHS
    0.06
    िछ
    0.06
    (pDX
    0.06
    _rhs
    0.06
    .createQuery
    0.06
     ruh
    0.06
    Act Density 0.268%

    No Known Activations