© 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. 10-GEMMASCOPE-RES-131K
    7. 9209
    Prev
    Next
    INDEX
    Explanations

    proper nouns and classified terms

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

    This neuron is tuned to pick out mid-word subword chunks—multi-letter sequences that serve as stems or affixes in proper names (e.g. “svang-,” “irai,” “ania”) and technical terms (e.g. “octa-,” “decy-,” “uril”).

    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
    艄
    -3.33
     zoude
    -2.77
    bluza
    -2.61
    歺
    -2.58
    犼
    -2.58
    轳
    -2.52
    鼍
    -2.50
     mily
    -2.48
     rése
    -2.48
     และ
    -2.47
    POSITIVE LOGITS
    There
    2.95
    They
    2.72
    One
    2.66
    Two
    2.64
    A
    2.63
    Only
    2.56
    Because
    2.55
    After
    2.50
    At
    2.48
    Not
    2.47
    Activations Density 0.175%

    No Known Activations