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

    Neuronpedia

    INDEX
    Explanations

    "

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on code-style identifier tokens—especially those containing or ending with “id” (e.g. typeid, offset).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Jos
    -0.07
     Taş
    -0.06
     notices
    -0.06
     Jog
    -0.06
     &)
    -0.06
     notably
    -0.06
    ถนน
    -0.06
    .Protocol
    -0.06
     次
    -0.06
     Hats
    -0.06
    POSITIVE LOGITS
    SO
    0.07
    _HERE
    0.06
    0.06
     Capt
    0.06
    lesia
    0.06
    olkata
    0.06
     "+↵
    0.06
    ázd
    0.06
    .geometry
    0.06
     natur
    0.06
    Act Density 0.206%

    No Known Activations