© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128303
    Prev
    Next
    INDEX
    Explanations

    The neuron activates on numeric tokens (numbers and decimals), essentially finding numerical expressions in the text.

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

    materials and devices

    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
    FilterChain
    -0.87
    不少
    -0.84
     but
    -0.83
    ātu
    -0.83
     квар
    -0.82
    ՝
    -0.81
     his
    -0.80
    tizens
    -0.79
    に行ってきました
    -0.79
     no
    -0.77
    POSITIVE LOGITS
    時間は
    0.92
    .>
    0.91
     装饰
    0.90
     ess
    0.88
    .«
    0.88
     scientific
    0.88
    认同
    0.87
    |}{$
    0.87
     automatica
    0.87
     gare
    0.87
    Activations Density 0.002%

    No Known Activations