© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 25539
    Prev
    Next
    INDEX
    Explanations

    Geography sections

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

    The neuron fires strongly on numeric and structured data in Wikipedia‐style articles—section headings (like “History” or “Geography”), numbers (years, population figures, coordinates), and infobox‐style entries.

    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
     campesinos
    -1.58
    谩
    -1.44
     campes
    -1.33
     Reſ
    -1.30
     саны
    -1.27
     narzęd
    -1.26
    čenje
    -1.26
    Referències
    -1.26
     auraient
    -1.25
     oceano
    -1.23
    POSITIVE LOGITS
     on
    1.37
     It
    1.31
     municipal
    1.30
    ⊞
    1.25
     it
    1.25
    ʽ
    1.23
     السياس
    1.21
    吸血
    1.21
     ступен
    1.20
     пли
    1.18
    Activations Density 0.097%

    No Known Activations