© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72314
    Prev
    Next
    INDEX
    Explanations

    code prefixes and identifiers

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

    The neuron activates on code import/package declarations—sequences of dot-separated identifiers (e.g. “com.utils.DBUtil”) followed by a semicolon.

    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
     Once
    -1.28
     Both
    -1.26
     nestled
    -1.15
     for
    -1.12
    eliminar
    -1.11
     bluse
    -1.10
     tremendously
    -1.09
     Obviously
    -1.09
     Regardless
    -1.09
     exquisite
    -1.08
    POSITIVE LOGITS
     vám
    1.36
     konkuren
    1.27
     rapporter
    1.23
     tyto
    1.23
     kullanıcı
    1.21
     filozof
    1.21
     vás
    1.20
     réjou
    1.16
     ulub
    1.16
     verhind
    1.15
    Activations Density 0.003%

    No Known Activations