© 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. 130918
    Prev
    Next
    INDEX
    Explanations

    The neuron activates on numeric tokens—integers, decimals, and numeric operators—i.e. any standalone numbers or number‐related symbols.

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

    repeated multiplication and exponents

    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
     Dumnezeu
    -0.97
     бампер
    -0.92
     τους
    -0.88
     ceva
    -0.85
    gegeven
    -0.83
     their
    -0.83
    륭
    -0.83
    袤
    -0.82
     gekle
    -0.82
    各类
    -0.81
    POSITIVE LOGITS
    Déf
    1.12
    LOAD
    0.96
    kjø
    0.93
     çocuklar
    0.93
    teinte
    0.93
     phénom
    0.92
     вспо
    0.92
     pelatih
    0.92
     Gastgeber
    0.91
    FRIEND
    0.91
    Activations Density 0.019%

    No Known Activations