© 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. 58472
    Prev
    Next
    INDEX
    Explanations

    code and comments

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

    The neuron fires on very high‐frequency/common tokens (e.g. frequent stopwords like “the,” “a,” common identifiers like “err” or keywords like “return,” and ubiquitous comment markers like “///”).

    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
     there
    -0.77
     fourteen
    -0.70
    Gente
    -0.69
    ܦ
    -0.69
     here
    -0.69
     löytyy
    -0.69
     Fonction
    -0.68
     eight
    -0.68
     canst
    -0.68
    Watt
    -0.67
    POSITIVE LOGITS
    (((
    0.74
     Aktien
    0.73
    broad
    0.73
    HYDR
    0.72
     الثلاث
    0.72
     reverting
    0.71
     знает
    0.69
    _));
    0.69
    スカ
    0.68
    AYS
    0.68
    Activations Density 0.037%

    No Known Activations