© 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. 10-GEMMASCOPE-RES-131K
    7. 86957
    Prev
    Next
    INDEX
    Explanations

    programming keywords before underscore

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

    The neuron fires on code‐style identifiers that include underscores—i.e. fragments of snake_case or underscored names (static_cast, make_pair, allow_receive, etc.).

    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
     which
    -2.58
     have
    -2.53
     but
    -2.47
     who
    -2.41
     will
    -2.41
     would
    -2.34
     to
    -2.31
     that
    -2.28
     still
    -2.27
     just
    -2.20
    POSITIVE LOGITS
     verdo
    2.17
    
    1.97
     bilde
    1.94
     praktis
    1.90
     meub
    1.88
    Karakteristik
    1.85
     manier
    1.84
    blech
    1.84
     linke
    1.84
     parfü
    1.84
    Activations Density 0.019%

    No Known Activations