© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 10564
    Prev
    Next
    INDEX
    Explanations

    codes and identifiers

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

    structured data formats like code, technical documents, product specifications, or tabular information with repeated formatting patterns.

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

    The neuron fires most strongly on numbered reference or section markers (e.g. “[1]”, “1.”, “[2]”, etc.), i.e. citation or outline numbering tokens.

    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
    铪
    -2.09
     italiens
    -2.08
    驁
    -1.99
    Ikke
    -1.98
     etek
    -1.98
    鲽
    -1.97
    ⠀⠀⠀⠀
    -1.96
    脔
    -1.95
     constater
    -1.95
    쐈
    -1.95
    POSITIVE LOGITS
     in
    2.52
     it
    2.33
     At
    2.27
     In
    2.16
     No
    2.08
     To
    2.05
     
    1.98
     If
    1.93
     to
    1.88
     As
    1.88
    Activations Density 0.007%

    No Known Activations