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

    Neuronpedia

    INDEX
    Explanations

    how

    np_max-act · gemini-2.0-flash

    questions and responses related to general knowledge and assistance.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron specifically fires on the word “How” when it begins an assistant’s question (e.g. “How can I help you today?”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     commitment
    -0.07
    ::_
    -0.07
    Process
    -0.06
    	attack
    -0.06
     sprintf
    -0.06
    Mit
    -0.06
     καλύ
    -0.06
    вал
    -0.06
    _stylesheet
    -0.06
    нь
    -0.06
    POSITIVE LOGITS
     उद
    0.07
    /s
    0.07
     místa
    0.07
    /m
    0.06
     Ре
    0.06
    /f
    0.06
     However
    0.06
    (freq
    0.06
    /F
    0.06
     Серг
    0.06
    Act Density 0.018%

    No Known Activations