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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 21710
    Prev
    Next
    INDEX
    Explanations

    closing brackets and parens

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

    The neuron fires on runs of closing delimiters (parentheses, brackets or braces), i.e. the ends of deeply nested code or data structures.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     メニュー
    -1.79
    こんに
    -1.71
     ใน
    -1.70
    壜
    -1.67
    引用元
    -1.66
    ilera
    -1.65
    Exploring
    -1.62
    zert
    -1.59
    ጓ
    -1.59
     apabila
    -1.59
    POSITIVE LOGITS
     the
    1.93
     at
    1.85
    u
    1.83
    is
    1.70
    na
    1.66
    k
    1.63
    r
    1.61
    ):
    1.58
    h
    1.55
    ;
    1.55
    Activations Density 0.022%

    No Known Activations