© 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. 41611
    Prev
    Next
    INDEX
    Explanations

    commas followed by new ideas

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

    This neuron spotlights compound adjectives ending in relation-indicating suffixes such as “-driven,” “-specific,” or “-based.”

    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
    dónde
    -1.45
     paroisse
    -1.31
     where
    -1.28
     Melhor
    -1.28
    veja
    -1.27
    風に
    -1.27
     obok
    -1.21
    nokia
    -1.19
    espero
    -1.18
     помощью
    -1.16
    POSITIVE LOGITS
     you
    1.75
    text
    1.59
    mml
    1.56
    begin
    1.55
     to
    1.54
    either
    1.47
     aksesuar
    1.46
    bigr
    1.46
    el
    1.44
    setcounter
    1.44
    Activations Density 0.117%

    No Known Activations