© 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. 10-GEMMASCOPE-RES-131K
    7. 103955
    Prev
    Next
    INDEX
    Explanations

    code completion

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

    whitespace indentation at the beginning of lines, particularly in code blocks.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    This neuron detects numeric tokens (any sequence containing digits, including integers and decimals).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     blissful
    -2.81
     When
    -2.56
    when
    -2.48
     perplexing
    -2.36
     three
    -2.34
     More
    -2.33
     whining
    -2.33
     unrelenting
    -2.31
     sparkly
    -2.31
     Creating
    -2.30
    POSITIVE LOGITS
    </i>
    3.41
     Dinas
    2.61
    </h3>
    2.50
     gesamte
    2.38
     genaue
    2.33
    鴴
    2.31
    </em>
    2.25
    鼋
    2.25
     dritte
    2.20
    睃
    2.20
    Activations Density 0.148%

    No Known Activations