© 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. 128183
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on SQL identifier tokens that name key columns—especially foreign-key or ID fields (e.g. “fk,” “_id,” “Hotel_ID,” “UserId,” “RoleId,” “results_fk”).

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

    database ids and foreign keys

    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
    سجيلات
    -0.79
     and
    -0.79
    osovo
    -0.73
    Karakteristik
    -0.73
     not
    -0.72
    分裂
    -0.71
    ｪ
    -0.70
     his
    -0.69
     or
    -0.69
    follower
    -0.69
    POSITIVE LOGITS
     исследования
    0.93
    .''
    0.90
     ausger
    0.90
     intenta
    0.89
    ARG
    0.85
     acela
    0.85
    niczy
    0.85
    siella
    0.85
    huri
    0.84
    GLA
    0.84
    Activations Density 0.008%

    No Known Activations