© 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. 34-GEMMASCOPE-RES-131K
    7. 62851
    Prev
    Next
    INDEX
    Explanations

    drawing geometric construction

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

    The neuron is primarily activated by numeric references—both spelled‐out numbers (like “two”) and digits—i.e. any token that denotes a number.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    اتور
    -0.84
    tene
    -0.79
     Chromat
    -0.75
     Maren
    -0.74
    halte
    -0.74
    roidery
    -0.73
     checkers
    -0.73
    czną
    -0.73
    Rash
    -0.73
     volumetric
    -0.73
    POSITIVE LOGITS
     endpoint
    0.95
     construction
    0.92
     constrained
    0.90
     endpoints
    0.85
    gcp
    0.84
    construction
    0.82
    Construction
    0.80
    Distance
    0.79
     command
    0.77
    androidx
    0.77
    Activations Density 0.012%

    No Known Activations