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

    Neuronpedia

    INDEX
    Explanations

    quoted speech

    np_max-act · gemini-2.0-flash

    The neuron activates on special structural or metadata tokens that mark the start of a header (e.g. “<|start_header_id|>”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ιστή
    -0.07
    _IE
    -0.07
     ix
    -0.06
    quip
    -0.06
     �
    -0.06
    itra
    -0.06
     Arabia
    -0.06
    =v
    -0.06
    _leave
    -0.06
    alled
    -0.06
    POSITIVE LOGITS
    dbg
    0.07
    SCII
    0.07
     azalt
    0.07
     encuentra
    0.06
     derec
    0.06
    .Nil
    0.06
     fart
    0.06
     verschiedene
    0.06
     объек
    0.06
     Derm
    0.06
    Act Density 0.124%

    No Known Activations