© 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. 128541
    Prev
    Next
    INDEX
    Explanations

    This neuron activates on isolated alphanumeric identifiers—especially numbers and uppercase codes (e.g. section/chapter numbers, IDs like “CVE-2000-4559,” “US,” version codes, percentages).

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

    CVE IDs, patent numbers, legal case numbers

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     ходе
    -1.13
     начале
    -1.02
     búsquedas
    -1.00
     conexão
    -0.99
     trying
    -0.96
     actitudes
    -0.95
     идеально
    -0.93
    ])){
    -0.91
     eigentlich
    -0.91
     exigencias
    -0.90
    POSITIVE LOGITS
    onedDateTime
    1.07
    вер
    1.01
     для
    0.99
     sano
    0.99
    кл
    0.98
    х
    0.97
    uitton
    0.97
     один
    0.97
    مرار
    0.96
    aguer
    0.96
    Activations Density 0.053%

    No Known Activations