© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 40102
    Prev
    Next
    INDEX
    Explanations

    court documents, multilingual origins

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

    This neuron is strongly activated by digits in the header metadata—i.e. it detects numeric sequences (case numbers, page numbers, dates) 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
    madeus
    -1.30
     próprias
    -1.27
     truk
    -1.22
    licante
    -1.21
    ACTS
    -1.20
    Ꮋ
    -1.18
    `/
    -1.17
    PRNewswire
    -1.16
    传递
    -1.15
    esthetics
    -1.14
    POSITIVE LOGITS
     aliments
    1.12
    anthrene
    1.05
    んでる
    1.03
     отве
    1.01
     Если
    0.98
    ánto
    0.96
     almeno
    0.96
    Extrait
    0.95
     Quận
    0.94
    -\
    0.93
    Activations Density 0.019%

    No Known Activations