© 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. 104122
    Prev
    Next
    INDEX
    Explanations

    document section headers

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

    The neuron fires on section or field headings in structured docs (e.g. “Properties”, “Description”, “Contents”, “Links”, “Enum”).

    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
     is
    -1.41
     HttpNotFound
    -1.18
    !”
    -1.18
    一趟
    -1.13
     producti
    -1.09
     metic
    -1.09
     vario
    -1.08
     centaines
    -1.06
    щена
    -1.06
     psychi
    -1.05
    POSITIVE LOGITS
     wußte
    1.47
     último
    1.47
    </h2>
    1.45
     aisladas
    1.37
    excellents
    1.36
     fleste
    1.36
     naszego
    1.34
     recent
    1.33
    bilden
    1.32
     quienes
    1.31
    Activations Density 0.054%

    No Known Activations