© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 71761
    Prev
    Next
    INDEX
    Explanations

    code block endings

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

    The neuron strongly activates on loop‐related code structures (for, while, recursive calls, template “endfor”, etc.) and the braces or markers that open and close those blocks.

    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
     miejscowości
    -0.94
    🈶
    -0.92
    Răspunde
    -0.92
    Тарих
    -0.91
    meng
    -0.91
     Stro
    -0.90
    ammans
    -0.90
    même
    -0.89
     jums
    -0.89
     météoro
    -0.89
    POSITIVE LOGITS
     if
    0.84
    していた
    0.82
     Planting
    0.79
     exhaus
    0.77
    冱
    0.77
     from
    0.77
    Substrate
    0.76
     Substrate
    0.75
     apor
    0.75
    フィギュア
    0.75
    Activations Density 0.001%

    No Known Activations