© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 12641
    Prev
    Next
    INDEX
    Explanations

    alphanumeric character removal

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

    The neuron fires on passages specifying technical character‐set or formatting requirements—e.g. rules about ASCII, alphanumeric, length, or password constraints.

    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
    usias
    0.70
    ienne
    0.58
    umbo
    0.58
    ucha
    0.54
    rike
    0.54
     barrel
    0.52
     truncated
    0.52
    絨
    0.50
     సొ
    0.50
    uche
    0.50
    POSITIVE LOGITS
    isdigit
    0.74
    excludes
    0.68
    Whitespace
    0.68
    数字
    0.68
     digits
    0.67
     alphanumeric
    0.67
    numbers
    0.66
    (/[
    0.66
    exclude
    0.65
     numbers
    0.65
    Activations Density 0.035%

    No Known Activations