© 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. 10-GEMMASCOPE-RES-131K
    7. 73590
    Prev
    Next
    INDEX
    Explanations

    reserved words or space

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

    The neuron fires on occurrences of the words “reserve” or “reserved” (e.g. in function names or descriptions of reserved space).

    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
    geſ
    -2.83
    矬
    -2.22
    嫑
    -2.14
    
    -2.14
    牁
    -2.09
    颪
    -2.09
    dámské
    -2.08
    蟶
    -2.05
    gefunden
    -2.03
    犼
    -2.03
    POSITIVE LOGITS
     has
    2.58
    m
    2.44
    es
    2.25
    </sub>
    2.22
    </h3>
    2.22
     for
    2.19
    ?).
    2.17
    w
    2.16
    ––
    2.06
    o
    2.06
    Activations Density 0.025%

    No Known Activations