© 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. 28174
    Prev
    Next
    INDEX
    Explanations

    function calls and declarations

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

    The neuron primarily fires on programming‐language syntax tokens (keywords like “return” or “if,” braces, punctuation, and other code elements) rather than natural-language words.

    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
     debacle
    -1.03
     inhalt
    -1.02
     overly
    -1.00
     饼
    -0.98
     dilapidated
    -0.98
    sous
    -0.98
     condividere
    -0.97
    怕
    -0.93
     bénéficiaire
    -0.93
     sauver
    -0.92
    POSITIVE LOGITS
     or
    1.34
     but
    1.17
     and
    1.11
     one
    1.00
     by
    1.00
    ().
    0.98
     first
    0.96
     last
    0.94
     (
    0.93
     yesterday
    0.93
    Activations Density 0.013%

    No Known Activations