© 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. 6828
    Prev
    Next
    INDEX
    Explanations

    protocol entities

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

    This neuron fires on punctuation-only tokens that mark metadata or directives—specifically the semicolons ending import/use statements and the hash signs beginning comment blocks or headings.

    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
     Gerald
    -0.85
    ダーク
    -0.83
    ₈
    -0.77
    jerk
    -0.73
    ⃗
    -0.70
    فيلم
    -0.69
    нского
    -0.68
    figcaption
    -0.67
    YNAMICS
    -0.67
     Arabic
    -0.66
    POSITIVE LOGITS
    Kafka
    0.77
     Kafka
    0.73
     PACK
    0.72
    ៗ
    0.71
    ilim
    0.70
     Overhead
    0.70
     hostility
    0.67
    বা
    0.66
    followed
    0.65
    ட்டு
    0.65
    Activations Density 0.075%

    No Known Activations