© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 53991
    Prev
    Next
    INDEX
    Explanations

    Speaking after, Author, Published

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

    The neuron consistently fires on isolated punctuation or numeric tokens—e.g. apostrophes in possessives, standalone digits in version strings, and closing-parenthesis marks.

    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
    elif
    -0.81
    起身
    -0.81
    BuilderFactory
    -0.80
     PAH
    -0.79
     Lukaku
    -0.79
    '%
    -0.78
     Tables
    -0.78
    MEDIUM
    -0.77
    メイ
    -0.77
    gure
    -0.77
    POSITIVE LOGITS
     systemic
    0.84
    修復
    0.83
    kein
    0.81
    Netto
    0.81
    usnya
    0.81
     banjir
    0.80
    ohne
    0.77
    Каждый
    0.77
     privés
    0.77
    Один
    0.77
    Activations Density 0.007%

    No Known Activations