© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 77245
    Prev
    Next
    INDEX
    Explanations

    section headers or labels

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

    The neuron strongly activates on numeric tokens (i.e., sequences of digits such as version numbers, years, counts, or other standalone numbers).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    words commonly used in technical contexts, particularly in software documentation, code comments, and structured data formats.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    technical/document-structure and metadata markers (e.g., section IDs, file/version or path strings, domain/header lines, code/math notation), rather than ordinary prose content.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     raggi
    -0.74
    UE
    -0.72
     Wire
    -0.72
    的力量
    -0.72
     baweł
    -0.69
    不合
    -0.68
    Run
    -0.68
    Wire
    -0.68
     Dae
    -0.68
    estream
    -0.67
    POSITIVE LOGITS
     comparte
    0.76
    verde
    0.75
    Cancell
    0.66
    ナー
    0.65
    miti
    0.64
     unat
    0.64
     dayan
    0.62
    Kurz
    0.62
     orien
    0.61
    팸
    0.60
    Activations Density 0.076%

    No Known Activations