© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130839
    Prev
    Next
    INDEX
    Explanations

    This neuron detects high-level structural markers and function words—especially prepositions (in, of, by, on), colons, and other linking tokens—that signal formal references or citations in text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    locations and geography concepts

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     anivers
    -1.20
     fiscale
    -1.14
     poë
    -1.13
    ÉM
    -1.06
     entusias
    -1.06
    توضیحات
    -1.05
     bobina
    -1.05
    *********/
    -1.04
     dikes
    -1.04
    　　【
    -1.04
    POSITIVE LOGITS
     location
    1.45
     geographical
    1.35
     locations
    1.27
     geography
    1.26
     land
    1.25
     جغرافيا
    1.20
     гео
    1.20
     travel
    1.19
     to
    1.19
     geographic
    1.16
    Activations Density 0.128%

    No Known Activations