© 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. 34-GEMMASCOPE-RES-131K
    7. 78733
    Prev
    Next
    INDEX
    Explanations

    How to or disease

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

    news article formatting features such as paragraph breaks, dates, and citation indicators.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron selectively activates on tokens that name courts or parties, numeric identifiers (e.g. case or amount figures), parenthetical acronyms, and other formal legal‐style markers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Dr
    -0.90
     waves
    -0.81
     can
    -0.80
    ⲣ
    -0.78
     Trying
    -0.74
     eager
    -0.73
     double
    -0.73
     durch
    -0.73
    ageable
    -0.73
    Cllr
    -0.72
    POSITIVE LOGITS
    PREMI
    1.34
     PREMIUM
    1.31
     nguoi
    1.04
     Mr
    1.02
    Mr
    0.93
    ganan
    0.89
    rà
    0.86
    rde
    0.86
     pedest
    0.82
    valet
    0.82
    Activations Density 0.013%

    No Known Activations