© 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. 67408
    Prev
    Next
    INDEX
    Explanations

    programming and testing keywords

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

    This neuron strongly activates on leading indentation and code‐structure tokens (i.e. the whitespace and opening tokens at the start of code lines).

    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
     will
    -1.24
     would
    -1.23
    calaure
    -1.11
    cún
    -0.98
     if
    -0.98
    (){
    -0.97
    Wait
    -0.93
     could
    -0.91
    社会
    -0.91
     was
    -0.91
    POSITIVE LOGITS
     TAMBIÉN
    1.32
     TAMBÉM
    1.24
     utiliser
    1.22
     akurat
    1.06
     personalización
    1.05
    ьет
    1.04
     tía
    1.03
    場は
    1.02
     obé
    1.00
    anium
    0.99
    Activations Density 0.020%

    No Known Activations