© 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. 34-GEMMASCOPE-RES-131K
    7. 94158
    Prev
    Next
    INDEX
    Explanations

    android code com list

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

    The neuron strongly fires on code‐specific tokens—especially punctuation like “(” and “'”—i.e. it identifies programming syntax.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Groves
    -0.88
     داشتند
    -0.86
    Osob
    -0.79
    lň
    -0.78
     ilustracja
    -0.78
     HILLS
    -0.78
    🤴
    -0.77
     berbagi
    -0.77
    itzenden
    -0.77
     geral
    -0.76
    POSITIVE LOGITS
     Stol
    0.89
    <sup>
    0.86
    rá
    0.84
     must
    0.84
    mitteilung
    0.82
    rased
    0.80
     com
    0.78
    ⑲
    0.77
     acompañada
    0.75
     extra
    0.73
    Activations Density 0.001%

    No Known Activations