© 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. 131023
    Prev
    Next
    INDEX
    Explanations

    The neuron activates on numeric literals (numbers) in both math problems and code.

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

    equations with q

    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
    āja
    -0.90
    ougou
    -0.85
    igerung
    -0.85
    ਖ਼
    -0.83
    mically
    -0.82
    pruchs
    -0.81
    ?,?,
    -0.80
    värr
    -0.79
     anonim
    -0.79
    meniz
    -0.79
    POSITIVE LOGITS
    msk
    0.86
    organiser
    0.85
     kelebihan
    0.82
     aiuta
    0.81
     supplémentaires
    0.79
     excess
    0.79
    underland
    0.77
    excess
    0.76
    Under
    0.75
     Hunde
    0.75
    Activations Density 0.023%

    No Known Activations