© 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. 45304
    Prev
    Next
    INDEX
    Explanations

    Q&A, mixed languages

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

    The neuron strongly activates on French boilerplate metadata phrases (e.g. “Page créé par” and “Mise à jour par”), i.e. accented French words indicating page authorship or update info.

    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
     augmenter
    -1.84
     dentes
    -1.84
    équipements
    -1.83
     nadru
    -1.77
     chaves
    -1.77
    庒
    -1.73
     prolon
    -1.73
     tirs
    -1.72
     bied
    -1.71
    
    -1.71
    POSITIVE LOGITS
     most
    1.75
     The
    1.66
    вості
    1.63
    ecé
    1.62
     realizes
    1.60
     was
    1.59
     sebagainya
    1.52
    S
    1.47
     currently
    1.47
     while
    1.46
    Activations Density 0.131%

    No Known Activations