© 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. 70871
    Prev
    Next
    INDEX
    Explanations

    fruit followed by food

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

    syntax characters and symbols in programming code like brackets, semicolons, and separators.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron is primarily triggered by numeric tokens—i.e. literal numbers in code or 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
     also
    -1.38
     still
    -1.09
     before
    -1.03
     didnt
    -0.95
     $
    -0.89
     Also
    -0.88
     B
    -0.88
     doesnt
    -0.88
    It
    -0.87
     only
    -0.86
    POSITIVE LOGITS
     réparer
    1.11
    
    1.05
    噐
    1.04
    谖
    1.00
     reparar
    1.00
    ניים
    0.96
     広場
    0.96
     camiset
    0.96
     réparation
    0.95
    傒
    0.94
    Activations Density 0.018%

    No Known Activations