© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72523
    Prev
    Next
    INDEX
    Explanations

    US/Canadian provinces in addresses

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

    The neuron specifically responds to numeric tokens (e.g. years, page numbers, zip codes, measurement values) in the text.

    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
     that
    -1.95
     so
    -1.85
     as
    -1.70
     any
    -1.70
     like
    -1.65
     without
    -1.61
     said
    -1.59
     really
    -1.59
     great
    -1.59
     also
    -1.57
    POSITIVE LOGITS
     vola
    1.43
     pié
    1.41
     formid
    1.34
     immen
    1.34
    Ꭶ
    1.34
     impecable
    1.33
     Atención
    1.32
    鹟
    1.32
     récol
    1.30
     applau
    1.30
    Activations Density 0.031%

    No Known Activations