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

    Neuronpedia

    INDEX
    Explanations

    _S

    np_max-act · gemini-2.0-flash

    The neuron responds to alphanumeric tokens that mix digits with letters or punctuation—e.g. version numbers, measurement specs, and coded identifiers.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     dışarı
    -0.06
     thần
    -0.06
     spanking
    -0.06
     ma
    -0.06
     fishermen
    -0.06
    545
    -0.06
    	diff
    -0.06
    'I
    -0.06
    fout
    -0.06
    NX
    -0.06
    POSITIVE LOGITS
     oblasti
    0.07
     соці
    0.06
    ाइव
    0.06
     forty
    0.06
    ]}↵
    0.06
    LOPT
    0.06
    Ord
    0.06
     judgment
    0.06
    ارف
    0.06
    .contentOffset
    0.06
    Act Density 0.020%

    No Known Activations