© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 101718
    Prev
    Next
    INDEX
    Explanations

    section and chapter markers

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

    The neuron specializes in detecting numeric legal citations and section or clause numbering (e.g., “section 70.1382(a)(ii)”, “this title”, numbered paragraphs).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     nke
    -1.30
    isable
    -1.27
    owany
    -1.27
     Danh
    -1.25
     手工
    -1.24
    <^
    -1.23
    のかな
    -1.23
     
    -1.23
     传统
    -1.23
    æring
    -1.21
    POSITIVE LOGITS
    ми
    1.59
     zostanie
    1.56
     Italij
    1.52
     sportiva
    1.52
     gwiazd
    1.51
    uudet
    1.48
    *}[
    1.46
     mosso
    1.45
     आइटम
    1.42
     Denken
    1.41
    Activations Density 0.005%

    No Known Activations