© 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. 22-GEMMASCOPE-RES-131K
    7. 129439
    Prev
    Next
    INDEX
    Explanations

    The neuron activates on formal framework and outline keywords—abstract, high-level labels like “elements,” “model,” “recommendations,” “goals,” “outcomes,” etc., used to structure academic or professional text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    goals, rules, and plans

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     spôso
    -0.96
    Then
    -0.94
    的问题
    -0.93
    itra
    -0.93
    根据
    -0.90
     الده
    -0.90
    raszamy
    -0.90
     pores
    -0.90
     läh
    -0.89
     sauté
    -0.89
    POSITIVE LOGITS
     and
    1.02
     while
    1.02
     without
    1.00
    translation
    0.96
    さの
    0.95
    ⅕
    0.94
     включает
    0.93
     at
    0.88
    驀
    0.88
     acheteurs
    0.88
    Activations Density 0.063%

    No Known Activations