© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 15746
    Prev
    Next
    INDEX
    Explanations

    The neuron is highly sensitive to numeric tokens, activating on numbers and numerical data in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Cornish
    0.57
    案
    0.57
     begr
    0.55
     matlab
    0.54
    ంతా
    0.52
     złoż
    0.52
    袢
    0.51
     शिवसेने
    0.50
     rv
    0.50
     ["[
    0.50
    POSITIVE LOGITS
     countries
    2.41
    countries
    2.19
    Countries
    2.19
     देशों
    2.14
     Countries
    2.11
     países
    2.06
     країн
    1.94
     کشورهای
    1.93
    各国
    1.90
     странах
    1.89
    Activations Density 0.389%

    No Known Activations