© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 45704
    Prev
    Next
    INDEX
    Explanations

    conditional code blocks

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

    This neuron fires on template- or macro-style interpolation and conditional syntax—e.g. double-curly braces ({{…}}), “if” expressions, and makefile-style $(…) constructs.

    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
    ícias
    -1.26
     gezin
    -1.25
     histó
    -1.19
    vät
    -1.17
    鰨
    -1.14
     rembourse
    -1.13
    ittää
    -1.11
    NAVIGATION
    -1.11
     douleurs
    -1.11
    🕗
    -1.10
    POSITIVE LOGITS
     type
    2.56
     mode
    2.16
    type
    1.85
     there
    1.75
     use
    1.67
    TYPE
    1.58
    mode
    1.53
     style
    1.52
     Type
    1.48
    Type
    1.48
    Activations Density 0.019%

    No Known Activations