© 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. 22-GEMMASCOPE-RES-131K
    7. 39839
    Prev
    Next
    INDEX
    Explanations

    Japanese sentence endings

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

    The neuron fires on standalone technical or metadata tokens—especially uppercase keywords, acronyms, and other markup identifiers (e.g. SQL commands, error codes, license or header fields).

    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
     for
    -1.34
    !
    -1.25
     of
    -1.21
     według
    -1.14
     Mě
    -1.13
     kiedyś
    -1.09
     The
    -1.09
     his
    -1.09
     Pří
    -1.08
    ślę
    -1.08
    POSITIVE LOGITS
    がありますが
    2.00
    <bos>
    1.59
    Détails
    1.45
    もありました
    1.44
    quelize
    1.41
    もありますが
    1.41
    ですが
    1.31
    がありました
    1.29
    もあって
    1.29
     reutiliz
    1.27
    Activations Density 0.007%

    No Known Activations