© 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. 22-GEMMASCOPE-RES-131K
    7. 70812
    Prev
    Next
    INDEX
    Explanations

    return programming constructs

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

    The neuron fires on programming‐style code tokens (brackets, semicolons, include/typedef/struct/class keywords), i.e. it detects code syntax rather than plain prose.

    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
     Projektu
    -1.55
    März
    -1.29
    kuat
    -1.27
     baterie
    -1.27
    ritsar
    -1.24
     Meningkatkan
    -1.23
     kilometr
    -1.19
     konflikt
    -1.18
    läs
    -1.18
     Jika
    -1.18
    POSITIVE LOGITS
    Jeudi
    1.15
    Publié
    1.07
     ー
    1.06
    Vendredi
    1.06
    Détails
    1.04
     incluso
    1.04
    damping
    1.03
     непосредственно
    1.01
     INCLUDES
    1.01
     laguna
    1.01
    Activations Density 0.000%

    No Known Activations