© 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. 34-GEMMASCOPE-RES-131K
    7. 59909
    Prev
    Next
    INDEX
    Explanations

    data types and structures

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

    The neuron responds to numeric literals (tokens made up of digits), i.e. it activates on numbers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    jš
    -0.90
     Çok
    -0.80
     memahami
    -0.79
     serialization
    -0.75
     Ewig
    -0.75
    riebe
    -0.74
    賀状
    -0.74
    طار
    -0.73
    щу
    -0.73
    derabad
    -0.73
    POSITIVE LOGITS
     dtype
    2.69
    dtype
    2.25
    Requires
    0.82
    astype
    0.80
    dtypes
    0.77
     dn
    0.76
    Portland
    0.76
    DataType
    0.75
     Requ
    0.75
     Doyle
    0.74
    Activations Density 0.011%

    No Known Activations