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

    Neuronpedia

    INDEX
    Explanations

    Random text snippets

    np_max-act · gemini-2.0-flash

    The neuron identifies email-style quotation markers (the “>” characters and their multi‐level variants) used to denote quoted or forwarded text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     till
    -0.07
     smoking
    -0.06
    'il
    -0.06
     Skyl
    -0.06
     Day
    -0.06
     utter
    -0.06
    ill
    -0.06
     upbringing
    -0.06
     rooting
    -0.06
     launched
    -0.06
    POSITIVE LOGITS
     �
    0.06
    )").
    0.06
    овж
    0.06
    став
    0.06
     ((((
    0.06
    /lab
    0.06
    ΙΟΥ
    0.06
    �
    0.06
     바
    0.06
    ../../../../
    0.06
    Act Density 0.013%

    No Known Activations