© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34941
    Prev
    Next
    INDEX
    Explanations

    legal preamble or definition phrases

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

    This neuron detects formal legal or official document lead-in markers and headings (e.g. “No.,” “Whereas,” “And Now,” “Upon consideration,” “Ordered,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     предметы
    -1.21
     أجل
    -1.21
     sebagainya
    -1.19
    の前に
    -1.19
    дото
    -1.12
    gbarkeit
    -1.12
    źć
    -1.11
     передний
    -1.10
    iální
    -1.09
    Кто
    -1.07
    POSITIVE LOGITS
     we
    1.32
     that
    1.30
     in
    1.20
     according
    1.19
    AutoModerator
    1.18
     the
    1.11
    ソフトウェア
    1.08
    ,
    1.02
     if
    1.02
    商品は
    0.99
    Activations Density 0.032%

    No Known Activations