© 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. 16706
    Prev
    Next
    INDEX
    Explanations

    court case citations

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

    The neuron fires on structural “metadata” lines—things like section titles, legal‐citation headers (volume/page numbers and “v.”), code‐snippet delimiters, and similar document‐level headings rather than running text.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    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
     Ingeniería
    -0.97
     coronel
    -0.91
     Ciebie
    -0.90
     Şi
    -0.90
     concr
    -0.89
     världen
    -0.89
     inconsciente
    -0.88
     ilgi
    -0.88
     dritte
    -0.88
    چگونه
    -0.87
    POSITIVE LOGITS
     but
    1.29
    USION
    0.99
     and
    0.99
    中略
    0.98
     with
    0.97
     which
    0.93
     jų
    0.91
     ядер
    0.91
    вання
    0.90
    mişti
    0.87
    Activations Density 0.019%

    No Known Activations