© 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. 128975
    Prev
    Next
    INDEX
    Explanations

    This neuron detects numeric reference or citation markers enclosed in square brackets.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    names starting with various letters

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    gridColumn
    -1.33
     cento
    -1.27
     one
    -1.25
    ÍSTICAS
    -1.23
     and
    -1.23
     persa
    -1.23
     where
    -1.22
     regata
    -1.21
     museo
    -1.21
     which
    -1.20
    POSITIVE LOGITS
     SEÑ
    1.35
     mieste
    1.34
     CONCEP
    1.32
     DÍA
    1.31
     tieto
    1.30
     みた
    1.28
     obviamente
    1.26
     fournir
    1.26
     foarte
    1.23
     Юрий
    1.21
    Activations Density 0.017%

    No Known Activations