© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 68394
    Prev
    Next
    INDEX
    Explanations

    programming code structure

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

    The neuron fires on code‐metadata tokens that name AST node types—things like “BlockStatement,” “ClassDeclaration,” or “ParameterizedTypeReference.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ite
    -1.88
    ise
    -1.69
    ">—
    -1.67
    ium
    -1.65
    ão
    -1.65
     wavering
    -1.64
    ый
    -1.57
    sive
    -1.57
    ые
    -1.55
    Both
    -1.53
    POSITIVE LOGITS
    ſſen
    1.65
    ですかね
    1.58
    ſſi
    1.57
     astonishing
    1.53
     extremely
    1.53
     素描
    1.48
     montagem
    1.45
     this
    1.44
     elegantly
    1.44
     have
    1.41
    Activations Density 0.001%

    No Known Activations