© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 112588
    Prev
    Next
    INDEX
    Explanations

    programming identifiers

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

    annotations or decorators in programming code, particularly those that mark important metadata like identifiers, parameters, paths, and query attributes.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron is strongly activated by programming‐language metadata annotations or attribute names (e.g. Param, DefaultValue, Path, Id, PrimaryKey, Key, Route, ModelAttribute, FromQuery).

    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
     from
    -1.58
     began
    -1.57
    さんも
    -1.55
     we
    -1.53
     with
    -1.48
    ):
    -1.46
     aumentado
    -1.46
    getMenuInflater
    -1.44
     for
    -1.40
     perdiendo
    -1.38
    POSITIVE LOGITS
     of
    2.11
     広場
    1.64
    
    1.53
     perché
    1.52
     magnific
    1.52
    AdvEx
    1.48
     ridiculously
    1.45
     SEVEN
    1.45
     artig
    1.41
     unſ
    1.41
    Activations Density 0.001%

    No Known Activations