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

    Neuronpedia

    INDEX
    Explanations

    where

    np_max-act · gemini-2.0-flash

    The neuron is looking for modal qualifier adverbs—most strongly firing on words like “possible” or “necessary.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    uters
    -0.07
    HTTPRequestOperation
    -0.06
     tragedies
    -0.06
    urpose
    -0.06
    fried
    -0.06
    .orange
    -0.06
    brief
    -0.06
    -graph
    -0.06
    rieben
    -0.06
     случай
    -0.06
    POSITIVE LOGITS
     Roma
    0.07
     살
    0.07
     unb
    0.06
    еров
    0.06
     Lond
    0.06
     hava
    0.06
    ****↵
    0.06
     Rarity
    0.06
     syncing
    0.06
    exter
    0.06
    Act Density 0.022%

    No Known Activations