© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128850
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on code keywords and imperative verbs that set or get properties—like “задать,” “отображение,” CSS property names (background‐color, border, border‐radius, opacity), and function names starting with “get.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    color, border, size

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    ִּ
    -1.34
    ּוֹ
    -1.33
     yyb
    -1.23
    afternoon
    -1.23
     kompati
    -1.21
     бампер
    -1.18
     каба
    -1.17
    paio
    -1.16
    setBold
    -1.16
     demografica
    -1.11
    POSITIVE LOGITS
     size
    1.32
     color
    1.20
     ecce
    1.09
     настройки
    1.03
     prefeitura
    1.02
    额外
    1.01
    ٧
    0.99
     colors
    0.95
    ه‌اند
    0.95
     опять
    0.94
    Activations Density 0.024%

    No Known Activations