© 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. 22-GEMMASCOPE-RES-131K
    7. 127938
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric tokens (integer or decimal literals) in the code.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    update statements

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     disfrute
    -1.05
    スタート
    -0.98
     épreuves
    -0.96
     audition
    -0.88
    rebuild
    -0.87
    TypeError
    -0.87
    restart
    -0.86
    volles
    -0.85
    gänger
    -0.85
    RectF
    -0.85
    POSITIVE LOGITS
     change
    1.33
     changing
    1.31
     changes
    1.24
     cambiar
    1.23
     changed
    1.17
     modifica
    1.05
    Geograf
    1.03
    改
    1.02
    เปลี่ยน
    1.01
     zmienia
    1.01
    Activations Density 0.005%

    No Known Activations