© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 103322
    Prev
    Next
    INDEX
    Explanations

    code delimiters and special characters

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

    This neuron detects shell‐script conditional tests—particularly the use of the “test” or “[” command with numeric comparison operators (e.g. -ne, -ge) and literal numbers.

    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
     (~
    -0.87
    Пря
    -0.85
    Nieder
    -0.84
    口气
    -0.83
    RIO
    -0.81
     Nieder
    -0.78
     ~(
    -0.78
     ад
    -0.78
    Шо
    -0.77
    ību
    -0.75
    POSITIVE LOGITS
     "";
    0.86
    $;
    0.85
    »;
    0.85
    $_,
    0.84
     '';
    0.83
    ”;
    0.81
     verpflich
    0.81
     ジャパン
    0.81
     ;
    0.81
     verpflichtet
    0.78
    Activations Density 0.049%

    No Known Activations