© 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. 34-GEMMASCOPE-RES-131K
    7. 33728
    Prev
    Next
    INDEX
    Explanations

    programming and technical text

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

    This neuron fires on code‐like syntax tokens—things such as comment delimiters (/**, */, *), angle‐bracket includes (<sstream>, <tuple>, etc.), hyphenated CSS/FX properties (-fx-…), and other punctuation-heavy code or annotation markers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     modifiers
    -0.73
     modific
    -0.70
     республи
    -0.69
     chan
    -0.69
    untamente
    -0.69
    ométrie
    -0.68
    天氣
    -0.68
    瑜伽
    -0.68
    कृति
    -0.68
    iii
    -0.67
    POSITIVE LOGITS
    人も
    0.75
    を見
    0.70
    晰
    0.66
    方も
    0.66
    favicon
    0.65
    clou
    0.65
    braith
    0.64
    銳
    0.63
     verke
    0.63
     indivíduo
    0.63
    Activations Density 0.055%

    No Known Activations