© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 32910
    Prev
    Next
    INDEX
    Explanations

    remainder or division by

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

    This neuron fires on tokens typical of code snippets, formulas, or other technical passages—e.g. function names (join), punctuation like “(” and “<”, and technical terms (encased).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     bénéfices
    -0.97
    Pág
    -0.95
    smh
    -0.91
     délais
    -0.90
     commentaires
    -0.90
    퐿
    -0.89
    comparable
    -0.89
     calendriers
    -0.87
     équipements
    -0.87
    tengo
    -0.87
    POSITIVE LOGITS
     in
    1.50
     from
    1.45
     by
    1.44
     of
    1.23
     on
    1.19
     where
    1.16
     at
    1.11
     what
    1.02
     for
    1.00
     their
    0.99
    Activations Density 0.119%

    No Known Activations