© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 127674
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on formal boilerplate phrasing and metadata labels in official documents—i.e. the prepositions, nouns, and structure words that make up standardized header or definitional lines.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408

    degree or of

    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
    Negative Logits
     toallas
    -1.06
     absorción
    -1.04
     razonable
    -1.04
    Would
    -1.02
     medición
    -0.99
     всі
    -0.98
     everyone
    -0.98
    
    -0.98
     locais
    -0.97
    ㍉
    -0.97
    POSITIVE LOGITS
     in
    1.26
     of
    1.16
    utado
    1.13
     rámci
    1.08
    TERNS
    1.07
    шње
    1.06
    of
    1.05
     décédé
    1.05
     schließt
    1.03
    else
    1.02
    Activations Density 0.032%

    No Known Activations