© 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. 22-GEMMASCOPE-RES-131K
    7. 72444
    Prev
    Next
    INDEX
    Explanations

    email closings and forwards

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

    The neuron strongly activates on email‐style separators and numeric tokens—i.e. runs of hyphens or strings of digits (dates, times, message IDs) in the header/footer.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    Sincerely
    -1.17
    женские
    -0.95
    Regards
    -0.93
    ﹡
    -0.86
    
    -0.84
    帏
    -0.83
    jspx
    -0.82
    дизайн
    -0.82
     Jør
    -0.81
    िने
    -0.81
    POSITIVE LOGITS
    lapsible
    0.88
     and
    0.87
     www
    0.81
    nship
    0.80
    のでしょうか
    0.79
     następujące
    0.79
    userRole
    0.79
     forgot
    0.77
    ypso
    0.77
     отве
    0.76
    Activations Density 0.024%

    No Known Activations