© 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. 34-GEMMASCOPE-RES-131K
    7. 107750
    Prev
    Next
    INDEX
    Explanations

    dar ra, igar ra

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

    The neuron responds to the structured, data‐heavy “box‐score” portions of sports/racing write-ups—i.e. tokens carrying numbers, codes, dates, odds, distances, and horse or jockey names in the result summary.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Side
    -0.77
     blended
    -0.73
     vů
    -0.73
    ınıza
    -0.71
    区間
    -0.70
     RMA
    -0.69
    みました
    -0.68
     wanted
    -0.68
     started
    -0.67
    bbl
    -0.65
    POSITIVE LOGITS
     COLLECTION
    0.80
     jumper
    0.80
     OLD
    0.79
    踊
    0.78
     offspring
    0.78
    藁
    0.78
     licensing
    0.75
    pref
    0.74
    COLLECTION
    0.73
    apore
    0.73
    Activations Density 0.009%

    No Known Activations