© 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. 84094
    Prev
    Next
    INDEX
    Explanations

    programming code declarations

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

    The neuron fires on code-level declarations and annotations—things like attribute or decorator syntax (e.g. the C# `[DllImport …]` and extern modifiers, GLSL `layout(...)`/`uniform` specs, Bazel `load(…)`, etc.) that introduce or configure definitions.

    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
     with
    -1.48
     only
    -1.14
     from
    -1.09
     if
    -1.04
    由于
    -0.96
     that
    -0.93
     each
    -0.93
    ذي
    -0.92
     consists
    -0.91
    :
    -0.90
    POSITIVE LOGITS
    Книга
    0.98
     понадоби
    0.97
    durante
    0.97
    ('');
    
    0.96
     ejemplos
    0.96
    apprendre
    0.96
    λλην
    0.96
    Жизнь
    0.96
    ㈰
    0.94
    भग
    0.94
    Activations Density 0.003%

    No Known Activations