© 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. 34-GEMMASCOPE-RES-131K
    7. 50300
    Prev
    Next
    INDEX
    Explanations

    Presidential, briefing, festival, proprietor

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

    The neuron strongly responds to formal, bureaucratic announcement language—especially capitalized titles, official categories, dates, and related tokens found in government or ceremonial notices.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    外国人
    -0.92
    魔力
    -0.91
     Zimbabwe
    -0.84
    谙
    -0.83
     Harare
    -0.83
     RVA
    -0.83
     Lankan
    -0.82
     vivas
    -0.82
     junt
    -0.81
    𝗎
    -0.81
    POSITIVE LOGITS
    HOC
    0.81
     Mechan
    0.79
    Gw
    0.77
     transversal
    0.76
    legends
    0.75
     Treasury
    0.75
    capac
    0.75
    ház
    0.74
     мэ
    0.74
    scm
    0.73
    Activations Density 0.018%

    No Known Activations