© 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. 72063
    Prev
    Next
    INDEX
    Explanations

    come-behind, haven, years, inc, failure, answer, time

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

    The main thing this neuron does is detect numerical expressions—especially decimal numbers indicating measurements or statistics.

    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
     about
    -2.13
     only
    -1.73
     would
    -1.63
     almost
    -1.59
    >
    -1.59
     very
    -1.56
     have
    -1.52
     a
    -1.47
    ).
    -1.44
    %
    -1.44
    POSITIVE LOGITS
     炖
    1.65
     jonge
    1.56
     trener
    1.55
     liefde
    1.52
     variados
    1.51
     zamanda
    1.50
     sezonu
    1.48
     salonu
    1.45
    such
    1.43
    🔉
    1.42
    Activations Density 0.125%

    No Known Activations