© 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. 22-GEMMASCOPE-RES-131K
    7. 36677
    Prev
    Next
    INDEX
    Explanations

    code delimiters; string literals

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

    The neuron fires most strongly on the module or file‐path tokens in import statements (e.g. the quoted path in a Go import or the class name plus terminating semicolon in a Java import).

    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
     by
    -1.03
     into
    -0.96
    ("."
    -0.93
     помочь
    -0.92
    ␤
    -0.92
     at
    -0.90
            
    -0.86
    続けて
    -0.83
     agrí
    -0.82
    狗
    -0.82
    POSITIVE LOGITS
     Auffassung
    1.02
    這些
    1.00
     Quais
    0.98
    seitige
    0.94
     tiež
    0.93
     sposob
    0.93
    これらの
    0.93
     veľmi
    0.92
    successful
    0.91
     tých
    0.89
    Activations Density 0.003%

    No Known Activations