© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 93338
    Prev
    Next
    INDEX
    Explanations

    programming keywords and code structures

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

    This neuron fires on heavily indented lines inside code (especially within function bodies), i.e. deeply nested or indented code blocks.

    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
     magasiner
    -1.37
     patienter
    -1.20
     concierto
    -1.10
     eksemp
    -1.10
     kamer
    -1.07
     gangen
    -1.07
     hennes
    -1.06
     mystérie
    -1.06
     tider
    -1.05
    員の
    -1.05
    POSITIVE LOGITS
     into
    1.05
    を実施
    1.00
     le
    0.98
    你是不是
    0.96
    if
    0.95
    äh
    0.94
    Odpo
    0.93
    Plusieurs
    0.93
     It
    0.91
    breviations
    0.90
    Activations Density 0.003%

    No Known Activations