© 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. 95640
    Prev
    Next
    INDEX
    Explanations

    company and place names

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

    The neuron activates on standalone numeric tokens (e.g. lines with just an integer), effectively spotting article‐length or metadata numbers.

    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
    cjon
    -1.24
     kopen
    -1.23
     quoc
    -1.14
     kazak
    -1.13
    Encyklopedia
    -1.13
     naam
    -1.09
     condomínio
    -1.09
     hant
    -1.09
    ВД
    -1.09
     ananas
    -1.07
    POSITIVE LOGITS
     and
    1.67
     both
    1.45
     that
    1.44
    Even
    1.34
    Both
    1.28
     with
    1.27
    We
    1.27
    Just
    1.26
    Two
    1.26
     which
    1.25
    Activations Density 0.001%

    No Known Activations