© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 113834
    Prev
    Next
    INDEX
    Explanations

    programming syntax and multi-lingual function calls

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

    line breaks or whitespace that separates numbered list items or code statements.

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

    The neuron activates on numeric tokens—digits, numbers, or numeric literals in the text.

    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
     one
    -2.55
     what
    -2.36
     after
    -2.34
    e
    -2.27
     while
    -2.20
     provide
    -2.11
     for
    -1.98
     three
    -1.94
     mostly
    -1.91
     two
    -1.88
    POSITIVE LOGITS
     siè
    2.02
     jezt
    2.02
    魠
    1.99
    臜
    1.88
     hubieran
    1.88
     spéciaux
    1.88
     professione
    1.88
     estábamos
    1.87
     montrent
    1.87
     variés
    1.87
    Activations Density 0.140%

    No Known Activations