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

    Neuronpedia

    INDEX
    Explanations

    avoiding detail or repetition

    np_max-act · gemini-2.0-flash

    The neuron activates on first‐person discourse markers or hedging/transition phrases (e.g. “I won’t go into…,” “I only want to add…”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Chrom
    -0.07
     export
    -0.07
     فراهم
    -0.07
     emails
    -0.06
    尼亚
    -0.06
    .bits
    -0.06
    .Match
    -0.06
     churches
    -0.06
    ив
    -0.06
    bra
    -0.06
    POSITIVE LOGITS
    _lvl
    0.07
    ("/{
    0.06
    =tk
    0.06
     ""),↵
    0.06
    	dataType
    0.06
    /al
    0.06
    .sale
    0.06
    ी)
    0.06
    ogh
    0.06
    >>>(
    0.06
    Act Density 0.013%

    No Known Activations