© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 8283
    Prev
    Next
    INDEX
    Explanations

    capitals of world cities

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

    The neuron fires on mentions of specific geographic locations (e.g. cities, neighborhoods, addresses).

    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
    陕西
    1.06
     Okanagan
    1.02
     Huntsville
    0.89
     Pensacola
    0.89
    岡山
    0.88
     towns
    0.86
     県
    0.85
    漳
    0.84
     Daytona
    0.84
    湖北
    0.84
    POSITIVE LOGITS
     서울
    1.42
    北京市
    1.36
    東京都
    1.29
     تهران
    1.27
    서울
    1.24
     রাজধানীর
    1.23
    서울특별시
    1.21
     東京都
    1.20
     दिल्ली
    1.13
    Metro
    1.05
    Activations Density 1.170%

    No Known Activations