© 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-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 104731
    Prev
    Next
    INDEX
    Explanations

    country, government, party, company, team

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

    The neuron detects genitive or possessive constructions—especially the ’s marker and entity nouns (e.g. country, government, party, province, firm) indicating ownership or belonging.

    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
     it
    -1.77
     in
    -1.59
     just
    -1.40
     of
    -1.36
     only
    -1.33
     all
    -1.26
     noted
    -1.22
    魔兽
    -1.16
    Their
    -1.15
     appropriately
    -1.14
    POSITIVE LOGITS
     jadx
    1.49
     همچنین
    1.47
     are
    1.47
     Atención
    1.45
    Varför
    1.33
    requipa
    1.30
     fabricación
    1.29
    檻
    1.28
    的人物
    1.28
     concernés
    1.27
    Activations Density 0.031%

    No Known Activations