© 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. 45248
    Prev
    Next
    INDEX
    Explanations

    SQL keywords and syntax

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

    The neuron selectively fires on numeric and code‐style tokens (e.g. multi-digit numbers and SQL/DDL keywords), effectively spotting numeric literals and code identifiers.

    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
     might
    -0.90
    体は
    -0.80
    끊
    -0.79
     seems
    -0.77
     getWidth
    -0.76
     to
    -0.76
     hoping
    -0.74
     aiming
    -0.71
     haven
    -0.71
     seemed
    -0.71
    POSITIVE LOGITS
    Kentucky
    1.09
     AGAINST
    0.99
    ően
    0.91
    PARATOR
    0.90
     MySQL
    0.90
     Kentucky
    0.87
    Kirche
    0.86
    CUR
    0.86
     STRA
    0.85
    DELIM
    0.84
    Activations Density 0.082%

    No Known Activations