© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 63829
    Prev
    Next
    INDEX
    Explanations

    node, grunt, watch

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

    The neuron strongly responds to programming‐style tokens—especially JavaScript/build‐tool keywords and symbols (e.g. “grunt,” “watch,” “using,” braces, assignment operators, etc.).

    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
     newBuilder
    -0.84
    PTT
    -0.83
     Larsen
    -0.81
     Chamberlain
    -0.81
    ✣
    -0.80
     Ramos
    -0.80
    ikistan
    -0.79
    reuth
    -0.79
    amenti
    -0.77
    亮的
    -0.77
    POSITIVE LOGITS
     Kentucky
    1.27
     Louisville
    1.23
    Kentucky
    1.13
    Louisville
    1.13
     grunt
    0.93
    Cincinnati
    0.87
    grunt
    0.85
     Cincinnati
    0.83
    Indianapolis
    0.81
     bourbon
    0.79
    Activations Density 0.005%

    No Known Activations