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

    Neuronpedia

    INDEX
    Explanations

    pinMode

    np_max-act · gemini-2.0-flash

    This neuron fires on the special “control” tokens used to mark the start/end of text and speaker‐header boundaries (e.g. <|begin_of_text|>, <|start_header_id|>, <|end_header_id|>, user/assistant tags).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ớ
    -0.07
    	body
    -0.06
     сон
    -0.06
    جوی
    -0.06
    教
    -0.06
    Manchester
    -0.06
     CM
    -0.06
     LW
    -0.06
     Swe
    -0.06
     Immediate
    -0.06
    POSITIVE LOGITS
     desperate
    0.07
     USART
    0.07
    .onViewCreated
    0.07
     beauty
    0.07
     Dön
    0.06
    (columns
    0.06
    .constants
    0.06
    inker
    0.06
    Mutable
    0.06
    χρι
    0.06
    Act Density 0.003%

    No Known Activations