© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 15315
    Prev
    Next
    INDEX
    Explanations

    closing brace or parentheses

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on code tokens that name or declare functions/methods (e.g. method names in function definitions).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ₩
    -0.96
    nalité
    -0.95
     उदाहरण
    -0.93
    いま
    -0.93
     szív
    -0.91
    laration
    -0.91
    🪥
    -0.91
    Так
    -0.88
    tså
    -0.88
     '+':
    -0.88
    POSITIVE LOGITS
     or
    1.03
     или
    1.00
     possibility
    0.99
    的多
    0.88
     is
    0.87
    頭皮
    0.85
    号码
    0.85
     ridden
    0.84
    <0xA3>
    0.80
     while
    0.79
    Activations Density 0.003%

    No Known Activations