© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128767
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on programming tokens that represent or contain numeric values (e.g. literals or identifiers with digits).

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

    code variables and definitions

    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
     signaler
    -1.16
     officier
    -1.11
     rente
    -1.10
     trous
    -1.06
     ivo
    -1.05
     boty
    -1.04
     magasiner
    -1.03
     dezelve
    -1.02
     vernis
    -1.02
     seigneur
    -1.02
    POSITIVE LOGITS
    </h1>
    1.27
     $
    1.13
     H
    1.11
     by
    1.08
     I
    1.07
    on
    1.01
     e
    1.00
    Q
    0.98
     A
    0.98
     Now
    0.95
    Activations Density 0.115%

    No Known Activations