© 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. 9766
    Prev
    Next
    INDEX
    Explanations

    code structure closing punctuation

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

    The neuron is detecting code punctuation and delimiter tokens (commas, semicolons, parentheses, braces, numeric literals) rather than actual words.

    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
     terminée
    -1.12
     imidlertid
    -1.02
     suivie
    -0.95
     kuitenkin
    -0.93
     装
    -0.93
    пьютер
    -0.93
     soudain
    -0.91
    нову
    -0.91
     Different
    -0.90
    графії
    -0.90
    POSITIVE LOGITS
     as
    1.13
     specifik
    1.04
     and
    1.04
     all
    1.02
     their
    0.98
     came
    0.97
    ıkları
    0.96
     furthering
    0.96
     primarily
    0.95
     sparked
    0.94
    Activations Density 0.034%

    No Known Activations