© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 9728
    Prev
    Next
    INDEX
    Explanations

    multi-lingual text fragments

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

    The neuron is picking out tokens from non-English (foreign) languages—i.e. words in other scripts or other languages embedded in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    擽
    -2.05
    ITY
    -1.68
    枼
    -1.65
    ſſel
    -1.57
     aiment
    -1.55
     في
    -1.52
    ISTIC
    -1.51
     nadal
    -1.51
     موب
    -1.48
    汩
    -1.48
    POSITIVE LOGITS
     
    1.77
     a
    1.77
    </h2>
    1.76
     '
    1.70
     an
    1.62
    啬
    1.62
    as
    1.55
     slated
    1.54
     large
    1.52
    '
    1.50
    Activations Density 0.001%

    No Known Activations