© 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. 98641
    Prev
    Next
    INDEX
    Explanations

    code blocks and control flow

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

    The neuron fires on tokens that close code blocks—e.g. “end” in Ruby/Lua and “}” in C-style languages.

    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
     dichter
    -1.23
     ,'
    -1.03
     aks
    -1.03
     niedersachsen
    -1.02
     pēc
    -0.97
    ⏟
    -0.97
     kleinere
    -0.97
    ﹠
    -0.96
    ChromeDriver
    -0.96
     Infatti
    -0.95
    POSITIVE LOGITS
     but
    1.08
     it
    1.05
     four
    0.93
    韬
    0.91
     before
    0.90
     would
    0.90
     more
    0.90
     time
    0.89
     any
    0.89
     all
    0.88
    Activations Density 0.002%

    No Known Activations