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

    Neuronpedia

    INDEX
    Explanations

    Factorial code

    np_max-act · gemini-2.0-flash

    The neuron fires on programming‐language tokens—especially include directives and other code keywords (e.g. library names or “program”)—marking spots in code.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    (players
    -0.07
     Permit
    -0.07
     بات
    -0.06
     backButton
    -0.06
    .Substring
    -0.06
     sexist
    -0.06
     občan
    -0.06
    ót
    -0.06
     dignity
    -0.06
     conferred
    -0.06
    POSITIVE LOGITS
    stitution
    0.07
    enville
    0.06
    unique
    0.06
    _svc
    0.06
     используют
    0.06
    _IMP
    0.06
    ń
    0.06
    {_
    0.06
    Configurer
    0.06
    .CSS
    0.06
    Act Density 0.010%

    No Known Activations