© 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. 71349
    Prev
    Next
    INDEX
    Explanations

    places and official titles

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

    The neuron fires on tokens typical of formal legal‐document headers—section titles, case or docket numbers, court names, and other all-caps or numbered headings.

    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
    𝑶
    -1.05
     surfer
    -1.05
    kori
    -1.04
    פיה
    -0.97
    ської
    -0.94
     Liberia
    -0.93
     poule
    -0.93
    vartis
    -0.93
     verbe
    -0.93
     男の子
    -0.93
    POSITIVE LOGITS
     didn
    1.27
     doesn
    1.14
     don
    1.13
    不像
    1.07
     individual
    1.05
    raphe
    1.02
     koš
    0.99
     have
    0.95
    </i>
    0.94
     children
    0.94
    Activations Density 0.081%

    No Known Activations