© 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. 58488
    Prev
    Next
    INDEX
    Explanations

    data retention, recovery, disaster

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

    The neuron fires on prominent technical nouns and keywords in documentation—especially “data” and related domain-specific terms like “disaster,” “recovery,” “retention,” and similar configuration or feature names.

    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
     russell
    -1.25
    isière
    -1.21
    ologische
    -1.16
     highschool
    -1.16
    ogi
    -1.15
    o
    -1.11
     Однако
    -1.11
    が届
    -1.10
    rivial
    -1.10
    Begriffsklä
    -1.09
    POSITIVE LOGITS
    🫴
    1.28
     its
    1.25
     last
    1.23
     mitos
    1.21
     it
    1.21
     persua
    1.17
    本章
    1.10
     our
    1.08
     inova
    1.08
    vegetables
    1.07
    Activations Density 0.114%

    No Known Activations