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

    Neuronpedia

    INDEX
    Explanations

    URLs or potential scams

    np_max-act · gemini-2.0-flash

    The neuron detects web address tokens, especially domain names and URL fragments (e.g. “.com,” “.biz,” “://www”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    анг
    -0.08
    .write
    -0.07
     shuffle
    -0.07
     Please
    -0.07
    .shuffle
    -0.07
    _with
    -0.06
     Chew
    -0.06
    enic
    -0.06
     grilled
    -0.06
     Breath
    -0.06
    POSITIVE LOGITS
     RECT
    0.07
     금액
    0.07
     dostan
    0.06
    $sql
    0.06
    ecedor
    0.06
    ayım
    0.06
    .Matrix
    0.06
    '];?>↵
    0.06
     wurde
    0.06
     ParameterDirection
    0.06
    Act Density 0.009%

    No Known Activations