© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 72910
    Prev
    Next
    INDEX
    Explanations

    programming keywords and syntax

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

    The neuron fires on programming‐language keywords and preprocessor tokens (e.g. C/C++ type specifiers like “unsigned,” include‐header names like “stdbool,” “utility>,” import strings like “fmt,” and “#define”).

    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
     their
    -1.65
     by
    -1.22
     most
    -1.21
     капю
    -1.17
     its
    -1.13
    iteln
    -1.13
    ජ
    -1.13
    uidado
    -1.12
     jesu
    -1.12
     your
    -1.11
    POSITIVE LOGITS
     nhung
    1.28
     aproximativ
    1.24
    approximately
    1.16
     básicamente
    1.14
     áng
    1.13
     также
    1.13
    ↵↵↵
    1.12
    
    1.12
     tambin
    1.10
    (>
    1.10
    Activations Density 0.000%

    No Known Activations