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

    Neuronpedia

    INDEX
    Explanations

    Encoded data

    np_max-act · gemini-2.0-flash

    The neuron detects long runs of Base64‐style text (as used in PEM/“BEGIN CERTIFICATE” blocks).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    emarks
    -0.06
    Frame
    -0.06
    938
    -0.06
    (test
    -0.06
     dest
    -0.06
    extras
    -0.06
    JUnit
    -0.06
     dak
    -0.06
     indir
    -0.06
     REST
    -0.06
    POSITIVE LOGITS
     fighting
    0.07
    -hidden
    0.07
    -double
    0.07
    ící
    0.06
    /social
    0.06
     staggering
    0.06
    -names
    0.06
    _fb
    0.06
    0.06
     автомати
    0.06
    Act Density 0.004%

    No Known Activations