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

    Neuronpedia

    INDEX
    Explanations

    Telling stories

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on Russian second-person imperative verbs (e.g. “расскажи,” “скажи”), i.e. commands asking the model to tell or explain something.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     criticised
    -0.07
    collapse
    -0.06
     bypass
    -0.06
    ceptor
    -0.06
     Property
    -0.06
     glare
    -0.06
    Apollo
    -0.06
     PUSH
    -0.06
     美国
    -0.06
    "P
    -0.06
    POSITIVE LOGITS
     recounted
    0.09
     рассказ
    0.09
     recounts
    0.09
     recount
    0.08
     него
    0.07
    ungs
    0.07
     розповід
    0.07
     latest
    0.06
    рати
    0.06
     fantast
    0.06
    Act Density 0.016%

    No Known Activations