© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2008
    Prev
    Next
    INDEX
    Explanations

    sections, options, limitations, trends

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

    This neuron fires on UI/navigation labels—words like “contact,” “reply,” “submission,” “helpdesk,” etc.—that indicate clickable links, buttons, or form controls.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    OutOf
    0.66
    among
    0.61
    ttf
    0.61
    🕰
    0.60
    kv
    0.59
    kk
    0.59
    CPF
    0.58
    cpf
    0.58
    kaggle
    0.57
    _(
    0.56
    POSITIVE LOGITS
     page
    0.97
     Guide
    0.88
     Page
    0.83
     section
    0.77
     pages
    0.77
     GUIDE
    0.77
     Pages
    0.76
     guide
    0.75
     Checklist
    0.75
     Analyzer
    0.75
    Activations Density 0.404%

    No Known Activations