© 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. 22-GEMMASCOPE-RES-131K
    4. 79793
    Prev
    Next
    INDEX
    Explanations

    health, community, separate, check

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

    comma-separated list items and enumerations in sentences.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron is selectively activating on content‐bearing nouns (often countable/plural “category” words like cells, forms, cases, items, processes, issues, etc.) that label entities or list items in technical or descriptive contexts.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     declarat
    -1.42
     venit
    -1.22
     NACIONAL
    -1.21
     statunit
    -1.18
     NSCoder
    -1.17
    เพราะ
    -1.11
     TableColumn
    -1.09
     applau
    -1.09
    Carex
    -1.09
     onPostExecute
    -1.08
    POSITIVE LOGITS
    It
    1.41
     Чтобы
    1.33
     ilustrasi
    1.24
    ようですが
    1.24
    dding
    1.23
     is
    1.23
     It
    1.20
    Ecotoxicity
    1.18
     alltså
    1.14
     kenyataan
    1.14
    Activations Density 0.426%

    No Known Activations