© 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. 22-GEMMASCOPE-RES-131K
    7. 104408
    Prev
    Next
    INDEX
    Explanations

    code, variable types, and "invention" phrases

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

    The neuron fires on programming‐language type‐ or mapping‐declaration syntax—e.g. generic angle brackets (“<”), mapping keywords, parentheses, and the “=>” arrow.

    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
    Procedimento
    -1.29
    США
    -1.12
     estampa
    -1.09
    Controllo
    -1.05
    获得了
    -1.05
    Halt
    -1.05
     abiye
    -1.04
    教练
    -1.03
    Paket
    -1.03
     العديد
    -1.02
    POSITIVE LOGITS
     of
    1.59
     године
    1.15
    deauna
    1.03
    ()]);
    1.02
     steder
    1.01
     shows
    0.99
    ///</
    0.98
     stories
    0.98
     locations
    0.97
    PARATUS
    0.96
    Activations Density 0.003%

    No Known Activations