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

    Neuronpedia

    INDEX
    Explanations

    complex sentences

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on commas (and nearby function words) when they mark out enumerated lists or itemized options.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     overnight
    -0.07
    -u
    -0.06
    最近
    -0.06
     También
    -0.06
    _Control
    -0.06
    dest
    -0.06
     �
    -0.06
    .paused
    -0.06
     shopper
    -0.06
    (storage
    -0.06
    POSITIVE LOGITS
    .Documents
    0.07
    vement
    0.07
    Ability
    0.07
    ους
    0.07
    Presentation
    0.07
     Speech
    0.07
    шается
    0.06
    difficulty
    0.06
    .protobuf
    0.06
    ptic
    0.06
    Act Density 0.373%

    No Known Activations