© 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. 10-GEMMASCOPE-RES-131K
    7. 20101
    Prev
    Next
    INDEX
    Explanations

    optic neuritis, MAVPACKED, MiR, FXML, packing paths

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

    The neuron fires on numeric literals (integers and floating‐point numbers), with activation roughly proportional to the value of the number.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     två
    -1.79
     moft
    -1.63
     quai
    -1.61
     kent
    -1.60
     まる
    -1.56
     !!!!
    -1.55
    approximately
    -1.54
    Hmmm
    -1.49
    ⁵
    -1.48
     FOUR
    -1.45
    POSITIVE LOGITS
     as
    2.92
     this
    2.63
     an
    2.08
     any
    1.99
     other
    1.95
     by
    1.92
     even
    1.90
     on
    1.84
     both
    1.83
     these
    1.76
    Activations Density 0.145%

    No Known Activations