© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 1392
    Prev
    Next
    INDEX
    Explanations

    The neuron primarily responds to unusual or out‐of‐vocabulary characters—especially odd punctuation or special Unicode symbols (e.g. fancy quotation marks and garbled bytes).

    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
    i
    0.57
    e
    0.55
    -
    0.52
    ry
    0.51
    o
    0.51
    bly
    0.49
    ‐
    0.49
    u
    0.48
    y
    0.48
    ien
    0.48
    POSITIVE LOGITS
     ошибка
    0.63
     因為
    0.63
     Frankly
    0.62
     作り
    0.62
    anjutnya
    0.61
     確認
    0.60
     заместитель
    0.60
     价
    0.59
     Agarwal
    0.58
     因为
    0.58
    Activations Density 0.000%

    No Known Activations