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

    Neuronpedia

    INDEX
    Explanations

    quotations

    np_max-act · gemini-2.0-flash

    The neuron primarily activates on terms that reference reports, studies, or authors (“report,” “paper,” “said,” “wrote,” etc.), i.e. markers of cited research or official findings.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     pageTitle
    -0.06
     Scotia
    -0.06
    (BitConverter
    -0.06
    (tool
    -0.06
    ibBundleOrNil
    -0.06
     тен
    -0.06
    Hat
    -0.06
    监
    -0.06
    GridColumn
    -0.06
    almö
    -0.06
    POSITIVE LOGITS
    acky
    0.07
     Volunteer
    0.07
    аний
    0.07
     estilo
    0.07
     Possible
    0.07
    produ
    0.07
    dict
    0.07
    才
    0.06
     destroy
    0.06
    azioni
    0.06
    Act Density 0.027%

    No Known Activations