© 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. 22-GEMMASCOPE-RES-131K
    4. 84129
    Prev
    Next
    INDEX
    Explanations

    programming keywords and identifiers

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

    The neuron detects programming identifiers—library/API calls, macros, class or function names, and other code‐specific tokens.

    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
    стит
    -0.96
     then
    -0.94
    $-
    -0.94
     toek
    -0.93
    atories
    -0.90
     will
    -0.90
     lebih
    -0.90
    💁
    -0.90
     plát
    -0.90
    ♣
    -0.90
    POSITIVE LOGITS
    おわりに
    0.99
    globin
    0.95
    変な
    0.88
     obeying
    0.85
    此处
    0.84
    可爱的
    0.84
     concorda
    0.82
     wrongfully
    0.81
    惯
    0.79
    ナナ
    0.79
    Activations Density 0.039%

    No Known Activations