© 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. 102096
    Prev
    Next
    INDEX
    Explanations

    Polska, Russian, Czech, Portuguese, problem, solve, possible

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

    The neuron flags word‐initial sub‐tokens in non-English (especially Slavic) languages—i.e. parts of words with diacritics or in Cyrillic, marking foreign‐language segments.

    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
    ↵
    -2.16
     when
    -1.80
     (
    -1.76
    </h2>
    -1.76
    <td>
    -1.73
    ;
    -1.69
     it
    -1.66
    :
    -1.61
    -
    -1.58
     as
    -1.53
    POSITIVE LOGITS
    鹩
    1.86
    1.82
    驁
    1.75
    泐
    1.73
    1.66
     parlar
    1.64
    する
    1.63
     ativo
    1.61
    ſſen
    1.58
    ܇
    1.58
    Activations Density 0.023%

    No Known Activations