© 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. 130779
    Prev
    Next
    INDEX
    Explanations

    The neuron fires most strongly on numeric literal tokens (numbers) in code.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    code blocks and punctuation

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     showing
    -1.09
     similar
    -1.02
     other
    -0.94
     Two
    -0.94
     any
    -0.93
     if
    -0.93
     those
    -0.92
     still
    -0.90
     Please
    -0.90
     not
    -0.90
    POSITIVE LOGITS
    很赞
    0.98
     HEN
    0.96
    ハンバーグ
    0.94
     Parfum
    0.94
     anta
    0.94
     spontan
    0.93
    quila
    0.93
     både
    0.93
     Conseils
    0.92
    
    0.92
    Activations Density 0.061%

    No Known Activations