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

    Neuronpedia

    INDEX
    Explanations

    to/should

    np_max-act · gemini-2.0-flash

    This neuron detects instructional or directive language—i.e. imperative verbs and phrases (like “have to,” “identify,” “capture,” “pick”) used to tell the user what actions to perform.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ChangedEventArgs
    -0.07
     contaminants
    -0.07
    .imp
    -0.07
     Pharmaceuticals
    -0.06
    (diff
    -0.06
    -sdk
    -0.06
     applicants
    -0.06
     benchmark
    -0.06
    Chip
    -0.06
     Grove
    -0.06
    POSITIVE LOGITS
     。↵
    0.06
     Anh
    0.06
     جلسه
    0.06
     �
    0.06
    究
    0.06
    .sell
    0.06
    지역
    0.06
     Nebraska
    0.06
    }/
    0.06
    _ATOMIC
    0.05
    Act Density 0.042%

    No Known Activations