© 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. 73785
    Prev
    Next
    INDEX
    Explanations

    code delimiters separating items

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

    The neuron is picking out numeric literal tokens (i.e. standalone numbers) in the text.

    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
     is
    -2.81
     But
    -2.73
     &
    -2.05
     D
    -2.00
     or
    -1.98
     K
    -1.98
     p
    -1.96
     All
    -1.94
     an
    -1.93
    ,
    -1.93
    POSITIVE LOGITS
    饜
    2.81
    
    2.56
    沵
    2.41
    삵
    2.41
     peuvent
    2.39
    齶
    2.34
     bloquer
    2.33
    鵙
    2.31
    ⑊
    2.31
    塒
    2.27
    Activations Density 0.021%

    No Known Activations