© 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. 67229
    Prev
    Next
    INDEX
    Explanations

    dates, months, and separators

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

    The neuron detects the numeric identifiers (and surrounding punctuation) that appear in press‐release metadata lines (e.g. the “/136” in “PRNewswire/136” or the “–” after “Marketwired”).

    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
     their
    -1.41
     both
    -1.17
     Krá
    -1.16
     any
    -1.12
    važ
    -1.09
     it
    -1.09
     for
    -1.06
     gwaran
    -1.05
     especially
    -1.04
     parar
    -1.03
    POSITIVE LOGITS
     firs
    1.12
     réessayer
    1.11
    ebilir
    1.10
     again
    1.10
     dépenses
    1.09
    utenants
    1.09
    Servus
    1.09
     ayur
    1.09
    ]=\
    1.07
    ulfon
    1.06
    Activations Density 0.000%

    No Known Activations