© 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. 22-GEMMASCOPE-RES-131K
    4. 67716
    Prev
    Next
    INDEX
    Explanations

    constructor or initializer calls

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

    The neuron is detecting object‐creation sites—spots where constructors or factory methods (e.g. “new …” or “…create(…)”) are invoked.

    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
     and
    -1.00
    Šalt
    -0.91
     facilmente
    -0.91
     秋冬
    -0.90
     triumphantly
    -0.90
     (?
    -0.88
    folie
    -0.87
     炒め
    -0.85
     Inoltre
    -0.84
     zowel
    -0.84
    POSITIVE LOGITS
     cavalli
    1.03
     before
    0.96
    olio
    0.96
    iseerd
    0.96
     any
    0.95
    łym
    0.90
     poważ
    0.90
     since
    0.87
    before
    0.86
    this
    0.85
    Activations Density 0.003%

    No Known Activations