© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34651
    Prev
    Next
    INDEX
    Explanations

    state and county legal contexts

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

    The neuron chiefly activates on long sequences of proper names and titles (e.g. judicial or official names with initials and honorifics) in court‐document headers.

    oai_token-act-pair · o4-miniTriggered 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
    Not in Any Lists

    No Comments

    Negative Logits
     sehr
    -1.43
     had
    -1.37
     idéale
    -1.33
    hộp
    -1.32
     détaillée
    -1.30
    Kedua
    -1.29
     recommandée
    -1.27
    )',
    -1.27
     from
    -1.26
    ၃
    -1.25
    POSITIVE LOGITS
     ofrece
    1.42
     obter
    1.36
     arriver
    1.33
     colossal
    1.32
    represents
    1.31
     usando
    1.30
     splendid
    1.29
     luscious
    1.29
     investir
    1.29
     investimento
    1.28
    Activations Density 0.098%

    No Known Activations