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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130353
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on capitalized single‐word names or acronyms—that is, proper nouns like Day, Mori, Anchor, FOCUS, Quartz, Manchester, loose, SIMON, etc.

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

    proper nouns and brands

    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
    Metode
    -1.93
    Entrée
    -1.71
     coraggio
    -1.70
    Spes
    -1.69
    雰囲
    -1.64
    ♪♪
    -1.64
    Ingredi
    -1.63
     工房
    -1.63
    kenal
    -1.63
    Réalisation
    -1.62
    POSITIVE LOGITS
     a
    2.09
    ad
    1.93
    ara
    1.91
    ap
    1.90
    ag
    1.88
    am
    1.84
     also
    1.80
    la
    1.79
    do
    1.78
    al
    1.77
    Activations Density 0.077%

    No Known Activations