© 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-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 102135
    Prev
    Next
    INDEX
    Explanations

    HTML tags and code

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

    text written in non-Latin scripts, particularly Hebrew, Arabic, Chinese, and other non-English languages.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron activates on tokens written in non-Latin scripts or containing nonstandard Latin characters (e.g. Hebrew, Arabic, accented letters).

    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
    -2.23
     to
    -1.93
    collier
    -1.73
     with
    -1.67
    acakt
    -1.66
     for
    -1.64
    bigli
    -1.61
     '
    -1.57
     on
    -1.55
    uniqlo
    -1.55
    POSITIVE LOGITS
     <=",
    2.30
    /
    
    2.22
    </h4>
    2.16
     <>",
    2.11
    /</
    2.09
     Lumpur
    2.06
     >=",
    2.05
    
    2.03
    PasswordEncoder
    2.03
    Ᏻ
    2.02
    Activations Density 0.001%

    No Known Activations