© 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. 24-GEMMASCOPE-2-RES-16K
    7. 7349
    Prev
    Next
    INDEX
    Explanations

    describing age groups

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

    The neuron fires strongly on the final subword chunks of various words—especially nominalizing or uncommon suffixes (e.g. “-ment,” “-ence(s),” “-akh,” “-podge,” “-iggledy,” “-ents”)—in effect marking word-ending fragments.

    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_24_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
     The
    0.95
     This
    0.86
     If
    0.86
    T
    0.85
     In
    0.83
     if
    0.81
    C
    0.77
    M
    0.77
    J
    0.76
    W
    0.76
    POSITIVE LOGITS
    মূলক
    0.85
     của
    0.73
    ային
    0.72
    льных
    0.71
     tenfold
    0.71
     bioactive
    0.69
     culturally
    0.68
     downwards
    0.68
    য়ের
    0.68
    ковый
    0.68
    Activations Density 0.000%

    No Known Activations