© 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. 67727
    Prev
    Next
    INDEX
    Explanations

    instances of a category

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

    The neuron consistently fires on large numeric tokens (e.g. four-digit counts like 972, 848, 828, etc.).

    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
    vensis
    -1.66
    了一個
    -1.61
    mäßigen
    -1.60
    ݮ
    -1.54
    atique
    -1.50
     any
    -1.48
    țe
    -1.45
     and
    -1.43
     willen
    -1.42
     хотят
    -1.42
    POSITIVE LOGITS
     of
    3.64
     have
    1.89
    が
    1.57
    では
    1.54
     are
    1.38
     companies
    1.35
     Klagen
    1.34
     (
    1.34
    は
    1.34
     has
    1.32
    Activations Density 0.033%

    No Known Activations