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

    Neuronpedia

    INDEX
    Explanations

    -am

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on capitalized function words that kick off new sentences (e.g. “The,” “In,” “For,” “It,” “One,” “Thus”), i.e. tokens at the start of sentences.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Ав
    -0.08
    VN
    -0.07
     PY
    -0.07
    ?).
    -0.06
     undead
    -0.06
     bitterness
    -0.06
    Az
    -0.06
    女性
    -0.06
     vend
    -0.06
    Mini
    -0.06
    POSITIVE LOGITS
    EntryPoint
    0.07
     Adjustment
    0.06
    _compiler
    0.06
    .getChild
    0.06
     presenting
    0.06
    ;",
    0.06
    -directed
    0.06
    }/#{
    0.06
    _clock
    0.06
    .boost
    0.06
    Act Density 0.068%

    No Known Activations