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

    Neuronpedia

    INDEX
    Explanations

    forum posts

    np_max-act · gemini-2.0-flash

    This neuron responds to the model’s internal dialogue‐management tokens and segment markers (e.g. <eot_id>, <start_header_id>, <end_header_id>), essentially detecting turn or segment boundaries.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .browser
    -0.07
    ительным
    -0.06
    итель
    -0.06
    CLK
    -0.06
     optimizing
    -0.06
    atitis
    -0.06
     soy
    -0.06
    КИ
    -0.06
     Toilet
    -0.06
     thẩm
    -0.06
    POSITIVE LOGITS
     FE
    0.06
     lesson
    0.06
    qu
    0.06
     mathematical
    0.06
     shooting
    0.06
     roads
    0.06
    unsafe
    0.06
     counting
    0.06
     rod
    0.06
    /e
    0.06
    Act Density 0.061%

    No Known Activations