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

    Neuronpedia

    INDEX
    Explanations

    code or math

    np_max-act · gemini-2.0-flash

    This neuron specifically detects the special header‐marker tokens (e.g. `<|start_header_id|>` and `<|end_header_id|>`) in the conversation format.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    azeera
    -0.07
     ترتیب
    -0.06
    .inputs
    -0.06
    _HP
    -0.06
     dragons
    -0.06
    -import
    -0.06
     manuscripts
    -0.06
     Braun
    -0.06
    -choice
    -0.05
    .Protocol
    -0.05
    POSITIVE LOGITS
     ung
    0.07
    petition
    0.07
     Sahara
    0.07
    ♪↵↵
    0.07
     tengo
    0.07
     후
    0.07
     واحدة
    0.07
    ώς
    0.06
    nya
    0.06
    NavItem
    0.06
    Act Density 0.072%

    No Known Activations