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

    Neuronpedia

    INDEX
    Explanations

    Interview transcripts

    np_max-act · gemini-2.0-flash

    The neuron detects mentions of people/agents and reporting actions—tokens that mark participants, speakers, or verbs introducing reported speech or descriptions.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
    _READY
    -0.07
    list
    -0.07
     fireplace
    -0.07
    Wild
    -0.07
     مسلمان
    -0.07
     tax
    -0.07
    Hash
    -0.07
     Accounting
    -0.06
     único
    -0.06
     robes
    -0.06
    POSITIVE LOGITS
     sunshine
    0.07
     котор
    0.06
     clam
    0.06
     overthrow
    0.06
     زیرا
    0.06
    	field
    0.05
     RID
    0.05
     Hospitals
    0.05
    /ay
    0.05
     onHide
    0.05
    Act Density 0.158%

    No Known Activations