© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 94018
    Prev
    Next
    INDEX
    Explanations

    Items after numbers and letters

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

    This neuron is triggered by code‐style tokens—especially Java package/import lines (fully qualified names like “com.tencent…”) and other structured, non-natural-language sequences.

    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
    istik
    -0.82
    mOn
    -0.80
     IMAGES
    -0.78
     motivates
    -0.78
    іод
    -0.77
     участь
    -0.76
    heses
    -0.74
     straff
    -0.74
     plantillas
    -0.73
     fiables
    -0.73
    POSITIVE LOGITS
     GRAM
    0.88
    irem
    0.84
    ianti
    0.83
    ært
    0.76
    getType
    0.75
    uyang
    0.75
    過ごし
    0.74
     закона
    0.73
    ENA
    0.73
     才
    0.72
    Activations Density 0.005%

    No Known Activations