© 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. 25896
    Prev
    Next
    INDEX
    Explanations

    chat message bubbles

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

    The neuron specializes in spotting code‐style syntax—particularly import statements, dot-separated API calls, semicolons, and similar programming tokens.

    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
     Title
    -0.81
    Mail
    -0.76
    Phone
    -0.75
    ctiva
    -0.73
     title
    -0.73
    NOW
    -0.72
    Title
    -0.71
    title
    -0.70
    หน
    -0.70
    あっ
    -0.69
    POSITIVE LOGITS
     bubble
    1.54
    bubble
    1.47
     messages
    1.41
    Bubble
    1.37
     bubbles
    1.37
     message
    1.36
     Bubble
    1.28
     sender
    1.17
     Bubbles
    1.09
    bubbles
    1.08
    Activations Density 0.017%

    No Known Activations