© 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. 10-GEMMASCOPE-RES-131K
    7. 38998
    Prev
    Next
    INDEX
    Explanations

    tensor decomposition and products

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

    The neuron activates on specialized mathematical jargon—especially advanced algebra and category-theory terms (e.g. “tensor,” “morphism,” “symmetric,” “graded,” “covariant derivative,” 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_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
     podczas
    -1.23
     privada
    -1.19
    攬
    -1.17
     it
    -1.16
     obsessed
    -1.12
    //
    -1.12
    Ингредиенты
    -1.09
     dadurch
    -1.07
     dalam
    -1.05
    CloseOperation
    -1.05
    POSITIVE LOGITS
    ݯ
    1.41
    ⛤
    1.36
    çons
    1.36
    ִים
    1.32
     ночи
    1.30
    ƍ
    1.29
    锜
    1.28
    itaires
    1.28
    billed
    1.27
    իլ
    1.25
    Activations Density 0.031%

    No Known Activations