© 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. 20529
    Prev
    Next
    INDEX
    Explanations

    French words and proper nouns

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

    The neuron fires on standalone specialized or proper‐noun tokens—e.g. names, headings, patent or legal matter labels, technical terms, biological taxa—basically rare domain‐specific nouns.

    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
     both
    -1.87
     more
    -1.73
     your
    -1.69
     all
    -1.65
     everything
    -1.63
     others
    -1.61
     will
    -1.58
     everyone
    -1.57
     our
    -1.57
    我们将
    -1.45
    POSITIVE LOGITS
     for
    1.41
     vernis
    1.37
     of
    1.30
     offrir
    1.30
     frein
    1.22
    了一种
    1.22
     certificat
    1.21
     eux
    1.20
    ge
    1.20
     pè
    1.16
    Activations Density 0.113%

    No Known Activations