© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 70772
    Prev
    Next
    INDEX
    Explanations

    foreign languages, English, technical and cultural terms

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find non-English words or tokens containing non-ASCII letters (e.g., Cyrillic or accented characters).

    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
     in
    -2.42
     to
    -2.20
    американ
    -1.67
    .
    -1.64
     on
    -1.55
     if
    -1.50
    </strong>
    -1.36
     internet
    -1.29
    存于
    -1.29
    essandro
    -1.27
    POSITIVE LOGITS
    quelize
    1.80
    笵
    1.80
    臵
    1.78
    tahui
    1.74
     berlina
    1.70
     our
    1.69
     doccia
    1.69
     qashqai
    1.68
    Además
    1.66
     perciò
    1.66
    Activations Density 0.021%

    No Known Activations