© 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. 34-GEMMASCOPE-RES-131K
    7. 78251
    Prev
    Next
    INDEX
    Explanations

    learning and memory deficit

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

    coding and technical documentation text elements including line breaks, language markers, and date formats.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron strongly fires on content‐bearing tokens—long or unusual words, symbols, and technical identifiers (e.g. library/class names, separator lines, hyphenated terms, numbers)—as opposed to common function words.

    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
    Sao
    -0.74
     constructor
    -0.71
    constructor
    -0.70
    constru
    -0.69
    trup
    -0.66
     Labrador
    -0.66
    QObject
    -0.66
     REG
    -0.65
    CEL
    -0.65
     Sao
    -0.64
    POSITIVE LOGITS
    infty
    0.81
    ndham
    0.71
    ləş
    0.69
    γρα
    0.68
    アップデート
    0.66
    enu
    0.66
    konen
    0.65
    aughey
    0.65
    roquia
    0.64
    کی
    0.63
    Activations Density 0.038%

    No Known Activations