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

    Neuronpedia

    INDEX
    Explanations

    technical writing

    np_max-act · gemini-2.0-flash

    The neuron is strongly activated by common English function words (e.g. articles, prepositions, conjunctions) rather than content words.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     whispers
    -0.07
    _contr
    -0.06
     acces
    -0.06
    _rest
    -0.06
     studio
    -0.06
    token
    -0.06
     coordinated
    -0.06
     Ferguson
    -0.06
     sais
    -0.06
    avax
    -0.06
    POSITIVE LOGITS
     OPT
    0.07
    .epoch
    0.07
    iał
    0.06
     Finland
    0.06
    874
    0.06
    _CBC
    0.06
     Election
    0.06
     перев
    0.06
    .IsNullOrEmpty
    0.06
    .code
    0.06
    Act Density 0.181%

    No Known Activations