© 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. 10-GEMMASCOPE-RES-131K
    4. 104691
    Prev
    Next
    INDEX
    Explanations

    email addresses and names

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

    colons that appear in metadata or structured header information, particularly in code comments indicating authors, filenames, dates, and other document properties.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    This neuron detects colon characters used as separators (e.g. in “Author:”, “year:”, “Source:”, or other key‐value labels) in code and documentation.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     extraordinaire
    -2.02
    見る
    -2.02
    𓆜
    -1.97
     принад
    -1.94
    荦
    -1.91
    ﹶ
    -1.84
     शुर
    -1.80
     및
    -1.80
    های
    -1.77
     vya
    -1.77
    POSITIVE LOGITS
     ancianos
    2.16
     are
    2.11
    </td>
    2.08
     reine
    2.05
    (
    2.02
    "$
    1.99
     "
    1.97
      
    1.93
     :
    1.84
     flirt
    1.80
    Activations Density 0.043%

    No Known Activations