© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 38342
    Prev
    Next
    INDEX
    Explanations

    quotation marks and following verbs

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

    This neuron fires on tokens that mark the start of new structural elements (like the first substantive word of a paragraph, list bullets or dashes, and section‐heading tokens).

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     a
    -2.47
    llten
    -1.95
     an
    -1.90
     mengucap
    -1.83
     and
    -1.83
     of
    -1.83
    ♪♪
    -1.64
     comportements
    -1.64
     that
    -1.63
    sich
    -1.63
    POSITIVE LOGITS
     needed
    1.91
     is
    1.84
     were
    1.83
     will
    1.80
     needs
    1.75
     are
    1.73
    嚕
    1.72
    を行います
    1.72
       
    1.70
    想了想
    1.70
    Activations Density 0.072%

    No Known Activations