© 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. 13343
    Prev
    Next
    INDEX
    Explanations

    programming variable names like num or n

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

    This neuron fires on variable identifiers (especially short, numeric‐related names like “n,” “num,” “r,” etc.) in code.

    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
     スカート
    -1.00
    costs
    -0.97
    𝒙
    -0.96
    without
    -0.95
    ctically
    -0.94
     beleid
    -0.93
    itiers
    -0.92
     PROFILE
    -0.92
    probably
    -0.91
     formes
    -0.88
    POSITIVE LOGITS
     number
    1.41
     integer
    1.07
     produces
    1.04
     will
    1.02
    ERON
    1.01
     ši
    0.96
     günlük
    0.96
    çant
    0.93
     років
    0.92
    gencias
    0.91
    Activations Density 0.025%

    No Known Activations