© 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. 34-GEMMASCOPE-RES-131K
    4. 505
    Prev
    Next
    INDEX
    Explanations

    Dragonball and Japan

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

    This neuron is most strongly activated by metadata tokens—standalone numbers (e.g. version or year figures) and title‐style words (like “Download” or “Game”) that appear in headings or labels.

    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
     駅
    -0.77
    ungszeit
    -0.71
    CustomerId
    -0.70
    Sask
    -0.67
    💽
    -0.66
    Pencil
    -0.66
    Carrot
    -0.66
     ‍
    -0.66
    صادر
    -0.65
    isinstance
    -0.65
    POSITIVE LOGITS
     hostess
    1.04
    桐
    0.86
     homeless
    0.85
     Homeless
    0.85
     karaoke
    0.81
     gai
    0.80
    Maj
    0.80
    rouvez
    0.79
    ikas
    0.79
     Cadillac
    0.78
    Activations Density 0.006%

    No Known Activations