© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34755
    Prev
    Next
    INDEX
    Explanations

    mathematical expressions and specific entities

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

    The neuron fires on parts of mathematical formulae—i.e. numbers, symbols (like “=”, “/”, “–”), variable labels (e.g. “S1”, “S2”), and parentheses—essentially marking tokens that are components of equations or series notation.

    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
     in
    -1.20
     бампер
    -1.12
     options
    -1.04
    ต้อง
    -1.01
    อีก
    -0.99
     Dostupné
    -0.99
    gemar
    -0.99
     new
    -0.96
     skog
    -0.96
     Umwelt
    -0.94
    POSITIVE LOGITS
     clô
    1.04
     ponieważ
    1.03
     dégust
    1.02
     extrait
    1.02
    лое
    1.01
     GRATU
    1.00
     alimenti
    0.99
     acrylique
    0.98
     valentín
    0.94
    を含
    0.93
    Activations Density 0.006%

    No Known Activations