© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 56775
    Prev
    Next
    INDEX
    Explanations

    Q&A, technical contexts

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

    The neuron fires on standalone code‐tokens: it lights up on programming keywords, symbols (like “–”), error words (“object”), SQL clauses (“WHERE”), compiler/configuration identifiers (“Visual”, “Grid”), copyright markers (“©”), and other non-natural-language tokens in code blocks.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     şe
    -0.87
     tragé
    -0.81
     siyah
    -0.80
     ресторан
    -0.77
     accepté
    -0.77
    apropri
    -0.77
     gümüş
    -0.75
     şir
    -0.74
     Prozesse
    -0.73
     народ
    -0.73
    POSITIVE LOGITS
    this
    1.08
     this
    1.00
     Watches
    0.85
    ỹ
    0.82
    лён
    0.82
     Woodruff
    0.82
    casts
    0.81
     bire
    0.80
    ugais
    0.79
     Newsletters
    0.77
    Activations Density 0.000%

    No Known Activations